949e5cfd-cb9c-48e7-a888-551981582a9b.txt

import os
import sys
with open(sys.argv[0]) as f:
    code = f.read() # read the code of this file ASAP, for logging
import uuid
import glob
import time
import contextlib
from dataclasses import dataclass

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torch.distributed as dist
import torch._inductor.config as config
from torch.nn.parallel import DistributedDataParallel as DDP
# Use of FlexAttention contributed by @KoszarskyB
from torch.nn.attention.flex_attention import flex_attention, create_block_mask
flex_attention = torch.compile(flex_attention, dynamic=False)
create_block_mask = torch.compile(create_block_mask, dynamic=False)

# -----------------------------------------------------------------------------
# Muon optimizer

def zeropower_via_svd(G, steps=None):
    U, S, V = G.svd()
    return U @ V.T

@torch.compile
def zeropower_via_newtonschulz5(G, steps=10, eps=1e-7):
    """
    Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a
    quintic iteration whose coefficients are selected to maximize the slope at zero. For the purpose
    of minimizing steps, it turns out to be empirically effective to keep increasing the slope at
    zero even beyond the point where the iteration no longer converges all the way to one everywhere
    on the interval. This iteration therefore does not produce UV^T but rather something like US'V^T
    where S' is diagonal with S_{ii}' ~ Uniform(0.5, 1.5), which turns out not to hurt model
    performance at all relative to UV^T, where USV^T = G is the SVD.
    """
    assert len(G.shape) == 2
    a, b, c = (3.4445, -4.7750,  2.0315)
    X = G.bfloat16()
    X /= (X.norm() + eps) # ensure top singular value <= 1
    if G.size(0) > G.size(1):
        X = X.T
    for _ in range(steps):
        A = X @ X.T
        B = b * A + c * A @ A # adapted from suggestion by @jxbz, @leloykun, and @YouJiacheng
        X = a * X + B @ X
    if G.size(0) > G.size(1):
        X = X.T
    return X

zeropower_backends = dict(svd=zeropower_via_svd, newtonschulz5=zeropower_via_newtonschulz5)

class Muon(torch.optim.Optimizer):
    """
    Muon - MomentUm Orthogonalized by Newton-schulz

    Muon internally runs standard SGD-momentum, and then performs an orthogonalization post-
    processing step, in which each 2D parameter's update is replaced with the nearest orthogonal
    matrix. To efficiently orthogonalize each update, we use a Newton-Schulz iteration, which has
    the advantage that it can be stably run in bfloat16 on the GPU.

    Some warnings:
    - This optimizer assumes that all parameters passed in are 2D.
    - It should not be used for the embedding layer, the final fully connected layer, or any {0,1}-D
    parameters; those should all be optimized by a standard method (e.g., AdamW).
    - To use it with 4D convolutional filters, it works well to just flatten their last 3 dimensions.
    - We believe it is unlikely to work well for training with small batch size.
    - We believe it may not work well for finetuning pretrained models, but we haven't tested this.
    - We have not yet tried this optimizer for training scenarios larger than NanoGPT (124M).

    Arguments:
        lr: The learning rate used by the internal SGD.
        momentum: The momentum used by the internal SGD.
        nesterov: Whether to use Nesterov-style momentum in the internal SGD. (recommended)
        backend: The chosen backend for the orthogonalization step. (recommended: 'newtonschulz5')
        backend_steps: The number of iteration steps to use in the backend, if it is iterative.
    """
    def __init__(self, params, lr=0.02, momentum=0.95, nesterov=True,
                 backend='newtonschulz5', backend_steps=5):
        defaults = dict(lr=lr, momentum=momentum, nesterov=nesterov, backend=backend, backend_steps=backend_steps)
        super().__init__(params, defaults)

    def step(self):

        for group in self.param_groups:

            lr = group['lr']
            momentum = group['momentum']
            zeropower_backend = zeropower_backends[group['backend']]

            # generate weight updates in distributed fashion
            total_params = sum(p.numel() for p in group['params'])
            updates_flat = torch.zeros(total_params, device='cuda', dtype=torch.bfloat16)
            curr_idx = 0
            for i, p in enumerate(group['params']):
                # luckily this will perfectly distribute a transformer with multiple of 4 layers to 8 GPUs
                if i % int(os.environ['WORLD_SIZE']) == int(os.environ['RANK']):
                    g = p.grad
                    assert g is not None
                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        state['momentum_buffer'] = torch.zeros_like(g)
                    buf = state['momentum_buffer']
                    buf.mul_(momentum).add_(g)
                    g = g.add(buf, alpha=momentum) if group['nesterov'] else buf
                    g = zeropower_backend(g, steps=group['backend_steps'])
                    g *= max(1, g.size(0)/g.size(1))**0.5
                    updates_flat[curr_idx:curr_idx+p.numel()] = g.flatten()
                curr_idx += p.numel()

            # sync updates across devices. we are not memory-constrained so can do this simple deserialization
            dist.all_reduce(updates_flat, op=dist.ReduceOp.SUM)

            # deserialize and apply updates
            curr_idx = 0
            for p in group['params']:
                g = updates_flat[curr_idx:curr_idx+p.numel()].view_as(p.data).type_as(p.data)
                p.data.add_(g, alpha=-lr)
                curr_idx += p.numel()

# -----------------------------------------------------------------------------
# PyTorch nn.Module definitions for the GPT-2 model

def norm(x):
    return F.rms_norm(x, (x.size(-1),))

class CastedLinear(nn.Linear):

    def __init__(self, in_features, out_features):
        super().__init__(in_features, out_features, bias=False)

    def forward(self, x):
        return F.linear(x, self.weight.to(x.dtype))

class Rotary(torch.nn.Module):

    def __init__(self, dim, base=10000):
        super().__init__()
        self.register_buffer('inv_freq', (1 / base) ** (torch.arange(0, dim, 2) / dim))
        self.seq_len_cached = None
        self.cos_cached = None
        self.sin_cached = None

    def forward(self, x):
        seq_len = x.shape[1]
        if seq_len != self.seq_len_cached:
            t = torch.arange(seq_len, device=x.device)
            freqs = torch.outer(t, self.inv_freq)
            self.seq_len_cached = seq_len
            self.cos_cached = freqs.cos()
            self.sin_cached = freqs.sin()
        cos, sin = self.cos_cached[None, :, None, :], self.sin_cached[None, :, None, :]
        # apply_rotary_emb(x, cos, sin)
        x1, x2 = x.chunk(2, dim=3)
        y1 = x1 * cos + x2 * sin
        y2 = x1 * (-sin) + x2 * cos
        return torch.cat((y1, y2), 3).type_as(x)

class CausalSelfAttention(nn.Module):

    def __init__(self, dim, n_head):
        super().__init__()
        assert dim % n_head == 0
        self.n_head = n_head
        self.c_q = CastedLinear(dim, dim)
        self.c_k = CastedLinear(dim, dim)
        self.c_v = CastedLinear(dim, dim)
        # value residual lambda
        self.lamb = nn.Parameter(torch.tensor(0.5)) # @Grad62304977
        # rotary embeddings
        self.rotary = Rotary(dim // n_head) # dim // n_head = head_dim
        # output projection
        self.c_proj = CastedLinear(dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x, vi, block_mask):
        B, T = x.size(0), x.size(1) # batch size, sequence length
        assert B == 1, "Must use batch size = 1 for FlexAttention"
        q = self.c_q(x).view(B, T, self.n_head, -1)
        k = self.c_k(x).view(B, T, self.n_head, -1)
        v = self.c_v(x).view(B, T, self.n_head, -1)
        v = (1 - self.lamb) * v + self.lamb * vi.view_as(v) # @Grad62304977
        q, k = norm(q), norm(k) # QK norm suggested by @Grad62304977
        q, k = self.rotary(q), self.rotary(k)
        y = flex_attention(q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2), block_mask=block_mask)
        y = y.transpose(1, 2).contiguous().view_as(x) # re-assemble all head outputs side by side
        y = self.c_proj(y)
        return y

class MLP(nn.Module):

    def __init__(self, dim):
        super().__init__()
        self.c_fc   = CastedLinear(dim, 4 * dim)
        self.c_proj = CastedLinear(4 * dim, dim)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x):
        x = self.c_fc(x)
        x = F.relu(x).square() # https://arxiv.org/abs/2109.08668v2; ~1-2% better than GELU; suggested by @SKYLINEZ007 and @Grad62304977
        x = self.c_proj(x)
        return x

class Block(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.attn = CausalSelfAttention(config.n_embd, config.n_head)
        self.mlp = MLP(config.n_embd)
        self.lambdas = nn.Parameter(torch.tensor([1., 0.]))

    def forward(self, x, vi, x0, block_mask):
        x = self.lambdas[0] * x + self.lambdas[1] * x0
        x = x + self.attn(norm(x), vi, block_mask)
        x = x + self.mlp(norm(x))
        return x

# -----------------------------------------------------------------------------
# The main GPT-2 model

@dataclass
class GPTConfig:
    vocab_size : int = 50304
    n_layer : int = 12
    n_head : int = 6 # head dim 128 suggested by @Grad62304977
    n_embd : int = 768

class GPT(nn.Module):

    def __init__(self, config):
        super().__init__()

        # U-net design by @brendanh0gan
        self.num_encoder_layers = config.n_layer // 2 # Half of the layers for encoder
        self.num_decoder_layers = config.n_layer - self.num_encoder_layers # Remaining for decoder
        # Add learnable skip connection weights for decoder layers
        self.skip_weights = nn.Parameter(torch.ones(self.num_decoder_layers))

        self.transformer = nn.ModuleDict(dict(
            wte = nn.Embedding(config.vocab_size, config.n_embd),
            # token value embeddings by @KoszarskyB - inspired by @Grad62304977's value residual learning
            vte = nn.Embedding(config.vocab_size, config.n_embd*12),
            h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
        ))
        self.lm_head = CastedLinear(config.n_embd, config.vocab_size)
        self.lm_head.weight.data.zero_() # @Grad62304977

    def forward(self, idx, target, attn_blocksize):

        docs = (idx == 50256).cumsum(0)
        def document_causal_mask(b, h, q_idx, kv_idx):
          causal_mask = q_idx >= kv_idx
          document_mask = docs[q_idx] == docs[kv_idx]
          window_mask = q_idx - kv_idx < attn_blocksize
          return causal_mask & document_mask & window_mask

        S = len(idx)
        block_mask = create_block_mask(document_causal_mask, None, None, S, S, device="cuda", _compile=True)

        # forward the GPT model itself
        x = self.transformer.wte(idx[None]) # token embeddings of shape (b, t, n_embd)
        x = norm(x) # @Grad62304977
        x0 = x
        vi = self.transformer.vte(idx[None]).chunk(12, dim=-1)

        # Store outputs for U-Net skip connections
        skip_connections = []
        # Encoder pass - process only the first half of the blocks
        for i in range(self.num_encoder_layers):
            x = self.transformer.h[i](x, vi[i], x0, block_mask)
            skip_connections.append(x)
        # Decoder pass - process the remaining blocks with weighted skip connections
        for i in range(self.num_decoder_layers):
            x = x + self.skip_weights[i] * skip_connections.pop()
            x = self.transformer.h[self.num_encoder_layers + i](x, vi[self.num_encoder_layers+i], x0, block_mask)

        x = norm(x)
        logits = self.lm_head(x)
        logits = 30 * torch.tanh(logits / 30) # @Grad62304977
        logits = logits.float()
        loss = F.cross_entropy(logits.view(-1, logits.size(-1)), target.view(-1))
        return loss

# -----------------------------------------------------------------------------
# Our own simple Distributed Data Loader

def _peek_data_shard(filename):
    # only reads the header, returns header data
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
    if header[0] != 20240520:
        print("ERROR: magic number mismatch in the data .bin file!")
        print("---> HINT: Are you passing in a correct file with --input_bin?")
        print("---> HINT: Dataset encoding changed recently, re-run data prepro or refer again to README")
        print("---> HINT: For example re-run: `python dev/data/tinyshakespeare.py`, then re-try")
        exit(1)
    assert header[1] == 1, "unsupported version"
    ntok = header[2] # number of tokens (claimed)
    return ntok # for now just return the number of tokens

def _load_data_shard(filename):
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
        assert header[0] == 20240520, "magic number mismatch in the data .bin file"
        assert header[1] == 1, "unsupported version"
        ntok = header[2] # number of tokens (claimed)
        # the rest of it are tokens, stored as uint16
        tokens = np.frombuffer(f.read(), dtype=np.uint16)
    assert len(tokens) == ntok, "number of tokens read does not match header?"
    return tokens

class DistributedDataLoader:
    def __init__(self, filename_pattern, T, process_rank, num_processes):
        self.process_rank = process_rank
        self.num_processes = num_processes
        self.T = T

        # glob files that match the pattern
        self.files = sorted(glob.glob(filename_pattern))
        assert len(self.files) > 0, f"did not find any files that match the pattern {filename_pattern}"

        # load and validate all data shards, count number of tokens in total
        ntok_total = 0
        for fname in self.files:
            shard_ntok = _peek_data_shard(fname)
            assert shard_ntok >= num_processes * T + 1
            ntok_total += int(shard_ntok)
        self.ntok_total = ntok_total

        self.reset()

    def reset(self):
        self.current_shard = -1
        self.advance()

    def advance(self): # advance to next data shard
        self.current_shard = (self.current_shard + 1) % len(self.files)
        self.current_position = self.process_rank * self.T
        self.tokens = _load_data_shard(self.files[self.current_shard])

    def next_batch(self):
        batch_size = self.T * self.num_processes
        buf = self.tokens[self.current_position:self.current_position+self.T+1]
        buf = torch.tensor(buf.astype(np.int32), dtype=torch.long)
        x = buf[:-1] # inputs
        y = buf[1:] # targets
        # advance current position and load next shard if necessary
        self.current_position += batch_size
        if self.current_position + batch_size >= len(self.tokens):
            self.advance()
        return x.cuda(), y.cuda()

# -----------------------------------------------------------------------------
# int main

@dataclass
class Hyperparameters:
    # data hyperparams
    input_bin : str = 'data/fineweb10B/fineweb_train_*.bin' # input .bin to train on
    input_val_bin : str = 'data/fineweb10B/fineweb_val_*.bin' # input .bin to eval validation loss on
    # optimization hyperparams
    batch_size : int = 8 # batch size, in sequences, across all devices
    sequence_length : int = 64*1024 # sequence length, in tokens
    num_iterations : int = 1530 # number of iterations to run
    warmup_iters : int = 0
    cooldown_iters : int = 600 # number of iterations of linear warmup/cooldown for triangular or trapezoidal schedule
    weight_decay : float = 0
    # evaluation and logging hyperparams
    val_loss_every : int = 125 # every how many steps to evaluate val loss? 0 for only at the end
    val_tokens : int = 10485760 # how many tokens of validation data? it's important to keep this fixed for consistent comparisons
    save_every : int = 0 # every how many steps to save the checkpoint? 0 for only at the end
args = Hyperparameters()

# set up DDP (distributed data parallel). torchrun sets this env variable
assert torch.cuda.is_available()
dist.init_process_group(backend='nccl')
ddp_rank = int(os.environ['RANK'])
ddp_local_rank = int(os.environ['LOCAL_RANK'])
ddp_world_size = int(os.environ['WORLD_SIZE'])
device = f'cuda:{ddp_local_rank}'
torch.cuda.set_device(device)
print(f"using device: {device}")
master_process = (ddp_rank == 0) # this process will do logging, checkpointing etc.

# begin logging
logfile = None
if master_process:
    run_id = str(uuid.uuid4())
    logdir = 'logs/%s/' % run_id
    os.makedirs(logdir, exist_ok=True)
    logfile = 'logs/%s.txt' % run_id
    # create the log file
    with open(logfile, "w") as f:
        # begin the log by printing this file (the Python code)
        f.write(code)
        f.write('='*100 + '\n')
def print0(s, logonly=False):
    if master_process:
        with open(logfile, "a") as f:
            if not logonly:
                print(s)
            f.write(s+'\n')
# log information about the hardware/software environment this is running on
# and print the full `nvidia-smi` to file
print0(f"Running pytorch {torch.version.__version__} compiled for CUDA {torch.version.cuda}\nnvidia-smi:")
import subprocess
result = subprocess.run(['nvidia-smi'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
print0(f'{result.stdout}', logonly=True)
print0('='*100, logonly=True)

# convenience variables
T = args.sequence_length
# calculate the number of steps to take in the val loop.
assert args.val_tokens % (T * ddp_world_size) == 0
val_steps = args.val_tokens // (T * ddp_world_size)
# calculate the steps of gradient accumulation required to attain the desired global batch size.
assert args.batch_size % (ddp_world_size) == 0
train_accumulation_steps = args.batch_size // ddp_world_size

# load tokens
train_loader = DistributedDataLoader(args.input_bin, T, ddp_rank, ddp_world_size)
val_loader = DistributedDataLoader(args.input_val_bin, T, ddp_rank, ddp_world_size)
print0(f"Training DataLoader: total number of tokens: {train_loader.ntok_total} across {len(train_loader.files)} files")
print0(f"Validation DataLoader: total number of tokens: {val_loader.ntok_total} across {len(val_loader.files)} files")
print0('='*100, logonly=True)
x, y = train_loader.next_batch()

# there are only 50257 unique GPT-2 tokens; we extend to nearest multiple of 128 for efficiency. suggested to me by @Grad62304977.
# this originates from Karpathy's experiments.
num_vocab = 50304
model = GPT(GPTConfig(vocab_size=num_vocab, n_layer=12, n_head=6, n_embd=768))
model = model.cuda().bfloat16()
for m in model.modules():
    if isinstance(m, CastedLinear):
        m.float()
if hasattr(config, "coordinate_descent_tuning"):
    config.coordinate_descent_tuning = True # suggested by @Chillee
model = torch.compile(model)
# here we wrap model into DDP container
model = DDP(model, device_ids=[ddp_local_rank])
raw_model = model.module # always contains the "raw" unwrapped model

# init the optimizer(s)
optimizer1 = torch.optim.Adam([raw_model.transformer.wte.weight, raw_model.transformer.vte.weight], lr=0.6, betas=(0.8, 0.95), fused=True)
optimizer2 = torch.optim.Adam([raw_model.lm_head.weight], lr=0.008, betas=(0.8, 0.95), fused=True)
params = list(raw_model.transformer.h.parameters())
matrix_params = [p for p in params if p.ndim == 2]
scalar_params = [p for p in params if p.ndim < 2] + [raw_model.skip_weights]
optimizer3 = Muon(matrix_params, lr=0.05, momentum=0.95)
optimizer4 = torch.optim.Adam(scalar_params, lr=0.04, betas=(0.8, 0.95), fused=True) # note that this learning rate is neither sensitive nor tuned
optimizers = [optimizer1, optimizer2, optimizer3, optimizer4]
# learning rate decay scheduler (linear warmup and cooldown)
def get_lr(it):
    assert it <= args.num_iterations
    # 1) linear warmup for warmup_iters steps
    if it < args.warmup_iters:
        return (it+1) / args.warmup_iters
    # 2) constant lr for a while
    elif it < args.num_iterations - args.cooldown_iters:
        return 1.0
    # 3) linear cooldown
    else:
        decay_ratio = (args.num_iterations - it) / args.cooldown_iters
        return decay_ratio
schedulers = [torch.optim.lr_scheduler.LambdaLR(opt, get_lr) for opt in optimizers]

# Start training loop
training_time_ms = 0
# start the clock
torch.cuda.synchronize()
t0 = time.time()
# begin training
for step in range(args.num_iterations + 1):
    last_step = (step == args.num_iterations)
    # This effectively ignores timing first 10 steps, which are slower for weird reasons.
    # Alternately, and slightly more correctly in terms of benchmarking, we could do 10
    # steps with dummy data first, and then re-initialize the model and reset the loader.
    if step == 10:
        training_time_ms = 0
        t0 = time.time()
    timed_steps = float('nan') if step <= 11 else (step - 10) + 1 # <= 11 to avoid bug in val

    # Set the attention blocksize for the current step, in chunks of 64. By @fernbear.bsky.social
    attn_blocksize = torch.tensor(64*((step/args.num_iterations * (1792 - 64) + 64)//64), dtype=torch.int, device='cuda')

    # once in a while evaluate the validation dataset
    if (last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # run validation batches
        model.eval()
        val_loader.reset()
        val_loss = 0.0
        for _ in range(val_steps):
            with torch.no_grad():
                x_val, y_val = val_loader.next_batch()
                val_loss += model(x_val, y_val, attn_blocksize=attn_blocksize)
        dist.all_reduce(val_loss, op=dist.ReduceOp.AVG)
        val_loss /= val_steps
        # log val loss to console and to logfile
        print0(f'step:{step}/{args.num_iterations} val_loss:{val_loss:.4f} train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms/(timed_steps-1):.2f}ms')
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    if master_process and (last_step or (args.save_every > 0 and step % args.save_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # save the state of the training process
        log = dict(step=step, code=code, model=raw_model.state_dict(), optimizers=[opt.state_dict() for opt in optimizers])
        torch.save(log, 'logs/%s/state_step%06d.pt' % (run_id, step))
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    # bit confusing: we want to make sure to eval on 0th iteration
    # but also after the very last iteration. so we loop for step <= num_iterations
    # instead of just < num_iterations (one extra due to <=), only to do
    # the validation/sampling one last time, and then we break right here as we're done.
    if last_step:
        break

    # --------------- TRAINING SECTION BEGIN -----------------
    model.train()
    for i in range(1, train_accumulation_steps+1):
        ctx = model.no_sync() if i < train_accumulation_steps else contextlib.nullcontext()
        with ctx: # there's no need to sync gradients every accumulation step
            # forward pass
            loss = model(x, y, attn_blocksize=attn_blocksize)
            # advance the dataset for the next batch
            x, y = train_loader.next_batch()
            # backward pass
            loss.backward()
        train_loss = loss.detach()
    for p in model.parameters():
        p.grad /= train_accumulation_steps
    # momentum warmup for Muon
    frac = min(step/300, 1)
    optimizer3.param_groups[0]['momentum'] = (1 - frac) * 0.85 + frac * 0.95
    # step the optimizers and schedulers
    for opt, sched in zip(optimizers, schedulers):
        opt.step()
        sched.step()
    # null the gradients
    model.zero_grad(set_to_none=True)
    # --------------- TRAINING SECTION END -------------------
    # everything that follows now is just diagnostics, prints, logging, etc.

    #dist.all_reduce(train_loss, op=dist.ReduceOp.AVG) # all-reducing the training loss would be more correct in terms of logging, but slower
    approx_time = training_time_ms + 1000 * (time.time() - t0)
    print0(f"step:{step+1}/{args.num_iterations} train_loss:{train_loss.item():.4f} train_time:{approx_time:.0f}ms step_avg:{approx_time/timed_steps:.2f}ms")

if master_process:
    print(f"peak memory consumption: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB")

# -------------------------------------------------------------------------
# clean up nice
dist.destroy_process_group()
====================================================================================================
Running pytorch 2.6.0.dev20241203+cu124 compiled for CUDA 12.4
nvidia-smi:
Thu Dec  5 04:25:28 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.06             Driver Version: 535.183.06   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA H100 80GB HBM3          On  | 00000000:19:00.0 Off |                    0 |
| N/A   38C    P0              75W / 700W |      3MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          On  | 00000000:3B:00.0 Off |                    0 |
| N/A   30C    P0              93W / 700W |     22MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          On  | 00000000:4C:00.0 Off |                    0 |
| N/A   31C    P0              98W / 700W |     22MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          On  | 00000000:5D:00.0 Off |                    0 |
| N/A   38C    P0             118W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          On  | 00000000:9B:00.0 Off |                    0 |
| N/A   38C    P0             123W / 700W |    529MiB / 81559MiB |      1%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          On  | 00000000:BB:00.0 Off |                    0 |
| N/A   29C    P0             110W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          On  | 00000000:CB:00.0 Off |                    0 |
| N/A   38C    P0             127W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          On  | 00000000:DB:00.0 Off |                    0 |
| N/A   30C    P0             118W / 700W |    529MiB / 81559MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
+---------------------------------------------------------------------------------------+

====================================================================================================
Training DataLoader: total number of tokens: 1100000000 across 11 files
Validation DataLoader: total number of tokens: 100000000 across 1 files
====================================================================================================
step:0/1530 val_loss:10.8258 train_time:0ms step_avg:nanms
step:1/1530 train_loss:10.8258 train_time:31490ms step_avg:nanms
step:2/1530 train_loss:10.0690 train_time:31601ms step_avg:nanms
step:3/1530 train_loss:8.3875 train_time:31763ms step_avg:nanms
step:4/1530 train_loss:7.5837 train_time:31922ms step_avg:nanms
step:5/1530 train_loss:7.4281 train_time:32081ms step_avg:nanms
step:6/1530 train_loss:6.9683 train_time:32242ms step_avg:nanms
step:7/1530 train_loss:7.1835 train_time:32402ms step_avg:nanms
step:8/1530 train_loss:6.7400 train_time:32564ms step_avg:nanms
step:9/1530 train_loss:6.6163 train_time:32724ms step_avg:nanms
step:10/1530 train_loss:6.5031 train_time:32885ms step_avg:nanms
step:11/1530 train_loss:6.4625 train_time:116ms step_avg:nanms
step:12/1530 train_loss:6.3588 train_time:276ms step_avg:nanms
step:13/1530 train_loss:6.2733 train_time:438ms step_avg:146.00ms
step:14/1530 train_loss:6.2395 train_time:598ms step_avg:149.50ms
step:15/1530 train_loss:6.1496 train_time:758ms step_avg:151.67ms
step:16/1530 train_loss:6.1067 train_time:919ms step_avg:153.15ms
step:17/1530 train_loss:6.1546 train_time:1079ms step_avg:154.09ms
step:18/1530 train_loss:5.9596 train_time:1239ms step_avg:154.93ms
step:19/1530 train_loss:5.9613 train_time:1400ms step_avg:155.56ms
step:20/1530 train_loss:5.6988 train_time:1561ms step_avg:156.06ms
step:21/1530 train_loss:5.9488 train_time:1721ms step_avg:156.47ms
step:22/1530 train_loss:6.1583 train_time:1881ms step_avg:156.72ms
step:23/1530 train_loss:5.8352 train_time:2041ms step_avg:157.02ms
step:24/1530 train_loss:6.0114 train_time:2202ms step_avg:157.27ms
step:25/1530 train_loss:5.6587 train_time:2362ms step_avg:157.49ms
step:26/1530 train_loss:5.5734 train_time:2523ms step_avg:157.68ms
step:27/1530 train_loss:5.7556 train_time:2683ms step_avg:157.82ms
step:28/1530 train_loss:5.3900 train_time:2842ms step_avg:157.92ms
step:29/1530 train_loss:5.6507 train_time:3003ms step_avg:158.04ms
step:30/1530 train_loss:5.4643 train_time:3163ms step_avg:158.13ms
step:31/1530 train_loss:5.4288 train_time:3323ms step_avg:158.24ms
step:32/1530 train_loss:5.2781 train_time:3483ms step_avg:158.31ms
step:33/1530 train_loss:5.5711 train_time:3643ms step_avg:158.38ms
step:34/1530 train_loss:5.4920 train_time:3803ms step_avg:158.48ms
step:35/1530 train_loss:5.6010 train_time:3963ms step_avg:158.54ms
step:36/1530 train_loss:5.5247 train_time:4123ms step_avg:158.59ms
step:37/1530 train_loss:5.4304 train_time:4283ms step_avg:158.64ms
step:38/1530 train_loss:5.2911 train_time:4444ms step_avg:158.72ms
step:39/1530 train_loss:5.3087 train_time:4604ms step_avg:158.77ms
step:40/1530 train_loss:5.2498 train_time:4764ms step_avg:158.81ms
step:41/1530 train_loss:5.2525 train_time:4925ms step_avg:158.86ms
step:42/1530 train_loss:5.1634 train_time:5085ms step_avg:158.91ms
step:43/1530 train_loss:5.2734 train_time:5245ms step_avg:158.94ms
step:44/1530 train_loss:5.2223 train_time:5408ms step_avg:159.06ms
step:45/1530 train_loss:5.3691 train_time:5569ms step_avg:159.12ms
step:46/1530 train_loss:5.1470 train_time:5730ms step_avg:159.15ms
step:47/1530 train_loss:5.0536 train_time:5889ms step_avg:159.16ms
step:48/1530 train_loss:5.1910 train_time:6050ms step_avg:159.22ms
step:49/1530 train_loss:5.1482 train_time:6211ms step_avg:159.24ms
step:50/1530 train_loss:5.2562 train_time:6370ms step_avg:159.25ms
step:51/1530 train_loss:5.1330 train_time:6531ms step_avg:159.29ms
step:52/1530 train_loss:5.0180 train_time:6691ms step_avg:159.31ms
step:53/1530 train_loss:5.1609 train_time:6851ms step_avg:159.32ms
step:54/1530 train_loss:5.0112 train_time:7011ms step_avg:159.35ms
step:55/1530 train_loss:5.4154 train_time:7172ms step_avg:159.38ms
step:56/1530 train_loss:5.0098 train_time:7333ms step_avg:159.42ms
step:57/1530 train_loss:4.8852 train_time:7492ms step_avg:159.41ms
step:58/1530 train_loss:5.0417 train_time:7653ms step_avg:159.44ms
step:59/1530 train_loss:5.0164 train_time:7814ms step_avg:159.47ms
step:60/1530 train_loss:5.1376 train_time:7975ms step_avg:159.49ms
step:61/1530 train_loss:4.8457 train_time:8137ms step_avg:159.54ms
step:62/1530 train_loss:4.9618 train_time:8297ms step_avg:159.57ms
step:63/1530 train_loss:4.9641 train_time:8457ms step_avg:159.57ms
step:64/1530 train_loss:4.8991 train_time:8618ms step_avg:159.59ms
step:65/1530 train_loss:4.7994 train_time:8777ms step_avg:159.59ms
step:66/1530 train_loss:4.9319 train_time:8939ms step_avg:159.62ms
step:67/1530 train_loss:4.8454 train_time:9099ms step_avg:159.63ms
step:68/1530 train_loss:5.0901 train_time:9259ms step_avg:159.65ms
step:69/1530 train_loss:4.7127 train_time:9420ms step_avg:159.66ms
step:70/1530 train_loss:4.8325 train_time:9579ms step_avg:159.65ms
step:71/1530 train_loss:4.9634 train_time:9739ms step_avg:159.66ms
step:72/1530 train_loss:4.8894 train_time:9900ms step_avg:159.68ms
step:73/1530 train_loss:4.7692 train_time:10060ms step_avg:159.68ms
step:74/1530 train_loss:4.9169 train_time:10220ms step_avg:159.69ms
step:75/1530 train_loss:4.8912 train_time:10380ms step_avg:159.69ms
step:76/1530 train_loss:4.8175 train_time:10540ms step_avg:159.70ms
step:77/1530 train_loss:4.9199 train_time:10700ms step_avg:159.70ms
step:78/1530 train_loss:5.1357 train_time:10860ms step_avg:159.71ms
step:79/1530 train_loss:4.8173 train_time:11021ms step_avg:159.72ms
step:80/1530 train_loss:4.8751 train_time:11181ms step_avg:159.72ms
step:81/1530 train_loss:4.6699 train_time:11341ms step_avg:159.73ms
step:82/1530 train_loss:4.8377 train_time:11502ms step_avg:159.75ms
step:83/1530 train_loss:4.7992 train_time:11663ms step_avg:159.76ms
step:84/1530 train_loss:4.7806 train_time:11823ms step_avg:159.77ms
step:85/1530 train_loss:4.6229 train_time:11983ms step_avg:159.77ms
step:86/1530 train_loss:4.8341 train_time:12143ms step_avg:159.78ms
step:87/1530 train_loss:4.7638 train_time:12303ms step_avg:159.78ms
step:88/1530 train_loss:4.7494 train_time:12463ms step_avg:159.78ms
step:89/1530 train_loss:4.7073 train_time:12624ms step_avg:159.80ms
step:90/1530 train_loss:4.6429 train_time:12783ms step_avg:159.79ms
step:91/1530 train_loss:4.6359 train_time:12943ms step_avg:159.79ms
step:92/1530 train_loss:4.8076 train_time:13104ms step_avg:159.80ms
step:93/1530 train_loss:4.6263 train_time:13264ms step_avg:159.81ms
step:94/1530 train_loss:4.6372 train_time:13424ms step_avg:159.81ms
step:95/1530 train_loss:4.6918 train_time:13585ms step_avg:159.82ms
step:96/1530 train_loss:4.5908 train_time:13745ms step_avg:159.82ms
step:97/1530 train_loss:4.6570 train_time:13905ms step_avg:159.82ms
step:98/1530 train_loss:4.5961 train_time:14065ms step_avg:159.83ms
step:99/1530 train_loss:4.6769 train_time:14227ms step_avg:159.85ms
step:100/1530 train_loss:4.6822 train_time:14387ms step_avg:159.86ms
step:101/1530 train_loss:4.5425 train_time:14547ms step_avg:159.86ms
step:102/1530 train_loss:4.7053 train_time:14708ms step_avg:159.87ms
step:103/1530 train_loss:4.5808 train_time:14869ms step_avg:159.88ms
step:104/1530 train_loss:4.5503 train_time:15029ms step_avg:159.89ms
step:105/1530 train_loss:4.5633 train_time:15188ms step_avg:159.87ms
step:106/1530 train_loss:4.6302 train_time:15349ms step_avg:159.89ms
step:107/1530 train_loss:4.5168 train_time:15510ms step_avg:159.89ms
step:108/1530 train_loss:4.3719 train_time:15670ms step_avg:159.90ms
step:109/1530 train_loss:4.4906 train_time:15831ms step_avg:159.91ms
step:110/1530 train_loss:4.4896 train_time:15991ms step_avg:159.91ms
step:111/1530 train_loss:4.4254 train_time:16151ms step_avg:159.91ms
step:112/1530 train_loss:4.5916 train_time:16310ms step_avg:159.90ms
step:113/1530 train_loss:4.4904 train_time:16471ms step_avg:159.91ms
step:114/1530 train_loss:4.3584 train_time:16631ms step_avg:159.91ms
step:115/1530 train_loss:4.5084 train_time:16794ms step_avg:159.94ms
step:116/1530 train_loss:4.4652 train_time:16958ms step_avg:159.98ms
step:117/1530 train_loss:4.3574 train_time:17122ms step_avg:160.02ms
step:118/1530 train_loss:4.5875 train_time:17285ms step_avg:160.04ms
step:119/1530 train_loss:4.4542 train_time:17448ms step_avg:160.07ms
step:120/1530 train_loss:4.3313 train_time:17612ms step_avg:160.11ms
step:121/1530 train_loss:4.2911 train_time:17777ms step_avg:160.15ms
step:122/1530 train_loss:4.4455 train_time:17941ms step_avg:160.19ms
step:123/1530 train_loss:4.2814 train_time:18104ms step_avg:160.22ms
step:124/1530 train_loss:4.5802 train_time:18267ms step_avg:160.24ms
step:125/1530 train_loss:4.4479 train_time:18433ms step_avg:160.28ms
step:125/1530 val_loss:4.4037 train_time:18479ms step_avg:160.69ms
step:126/1530 train_loss:4.4187 train_time:18597ms step_avg:160.32ms
step:127/1530 train_loss:4.4447 train_time:18763ms step_avg:160.37ms
step:128/1530 train_loss:4.3765 train_time:18927ms step_avg:160.40ms
step:129/1530 train_loss:4.6730 train_time:19090ms step_avg:160.42ms
step:130/1530 train_loss:4.3537 train_time:19255ms step_avg:160.46ms
step:131/1530 train_loss:4.4015 train_time:19418ms step_avg:160.48ms
step:132/1530 train_loss:4.3476 train_time:19581ms step_avg:160.50ms
step:133/1530 train_loss:4.4416 train_time:19745ms step_avg:160.53ms
step:134/1530 train_loss:4.2528 train_time:19909ms step_avg:160.56ms
step:135/1530 train_loss:4.4429 train_time:20075ms step_avg:160.60ms
step:136/1530 train_loss:4.2171 train_time:20239ms step_avg:160.62ms
step:137/1530 train_loss:4.3678 train_time:20402ms step_avg:160.65ms
step:138/1530 train_loss:4.2715 train_time:20565ms step_avg:160.66ms
step:139/1530 train_loss:4.3679 train_time:20730ms step_avg:160.69ms
step:140/1530 train_loss:4.4679 train_time:20894ms step_avg:160.72ms
step:141/1530 train_loss:4.3085 train_time:21058ms step_avg:160.75ms
step:142/1530 train_loss:4.3001 train_time:21221ms step_avg:160.76ms
step:143/1530 train_loss:4.2419 train_time:21384ms step_avg:160.79ms
step:144/1530 train_loss:4.3559 train_time:21549ms step_avg:160.82ms
step:145/1530 train_loss:4.3099 train_time:21714ms step_avg:160.84ms
step:146/1530 train_loss:4.1833 train_time:21878ms step_avg:160.87ms
step:147/1530 train_loss:4.3309 train_time:22041ms step_avg:160.88ms
step:148/1530 train_loss:4.3534 train_time:22205ms step_avg:160.90ms
step:149/1530 train_loss:4.2974 train_time:22370ms step_avg:160.93ms
step:150/1530 train_loss:4.4392 train_time:22533ms step_avg:160.95ms
step:151/1530 train_loss:4.2607 train_time:22698ms step_avg:160.98ms
step:152/1530 train_loss:4.2597 train_time:22861ms step_avg:160.99ms
step:153/1530 train_loss:4.3623 train_time:23024ms step_avg:161.01ms
step:154/1530 train_loss:4.3610 train_time:23188ms step_avg:161.03ms
step:155/1530 train_loss:4.2512 train_time:23353ms step_avg:161.05ms
step:156/1530 train_loss:4.3412 train_time:23516ms step_avg:161.07ms
step:157/1530 train_loss:4.4107 train_time:23681ms step_avg:161.09ms
step:158/1530 train_loss:4.2382 train_time:23844ms step_avg:161.11ms
step:159/1530 train_loss:4.3016 train_time:24007ms step_avg:161.12ms
step:160/1530 train_loss:4.1113 train_time:24173ms step_avg:161.15ms
step:161/1530 train_loss:4.3385 train_time:24337ms step_avg:161.17ms
step:162/1530 train_loss:4.3570 train_time:24500ms step_avg:161.19ms
step:163/1530 train_loss:4.3332 train_time:24664ms step_avg:161.20ms
step:164/1530 train_loss:4.1767 train_time:24827ms step_avg:161.22ms
step:165/1530 train_loss:4.2790 train_time:24991ms step_avg:161.23ms
step:166/1530 train_loss:4.3368 train_time:25155ms step_avg:161.25ms
step:167/1530 train_loss:4.1942 train_time:25318ms step_avg:161.26ms
step:168/1530 train_loss:4.2847 train_time:25481ms step_avg:161.27ms
step:169/1530 train_loss:4.1612 train_time:25645ms step_avg:161.29ms
step:170/1530 train_loss:4.0287 train_time:25810ms step_avg:161.31ms
step:171/1530 train_loss:4.2070 train_time:25974ms step_avg:161.33ms
step:172/1530 train_loss:4.2021 train_time:26137ms step_avg:161.34ms
step:173/1530 train_loss:4.2633 train_time:26301ms step_avg:161.35ms
step:174/1530 train_loss:4.4082 train_time:26464ms step_avg:161.36ms
step:175/1530 train_loss:4.2324 train_time:26626ms step_avg:161.37ms
step:176/1530 train_loss:4.0778 train_time:26788ms step_avg:161.38ms
step:177/1530 train_loss:4.0552 train_time:26951ms step_avg:161.38ms
step:178/1530 train_loss:4.1746 train_time:27114ms step_avg:161.39ms
step:179/1530 train_loss:4.1202 train_time:27278ms step_avg:161.41ms
step:180/1530 train_loss:4.1021 train_time:27440ms step_avg:161.41ms
step:181/1530 train_loss:4.2872 train_time:27602ms step_avg:161.42ms
step:182/1530 train_loss:4.1490 train_time:27766ms step_avg:161.43ms
step:183/1530 train_loss:4.1203 train_time:27927ms step_avg:161.43ms
step:184/1530 train_loss:4.1227 train_time:28092ms step_avg:161.45ms
step:185/1530 train_loss:4.1990 train_time:28255ms step_avg:161.46ms
step:186/1530 train_loss:4.1682 train_time:28417ms step_avg:161.46ms
step:187/1530 train_loss:4.2310 train_time:28580ms step_avg:161.47ms
step:188/1530 train_loss:4.1565 train_time:28877ms step_avg:162.23ms
step:189/1530 train_loss:4.1121 train_time:29207ms step_avg:163.16ms
step:190/1530 train_loss:4.2039 train_time:29369ms step_avg:163.16ms
step:191/1530 train_loss:4.0656 train_time:29531ms step_avg:163.16ms
step:192/1530 train_loss:4.0223 train_time:29695ms step_avg:163.16ms
step:193/1530 train_loss:4.2347 train_time:29857ms step_avg:163.15ms
step:194/1530 train_loss:4.1675 train_time:30019ms step_avg:163.15ms
step:195/1530 train_loss:4.3463 train_time:30182ms step_avg:163.15ms
step:196/1530 train_loss:4.1704 train_time:30346ms step_avg:163.15ms
step:197/1530 train_loss:4.0448 train_time:30509ms step_avg:163.15ms
step:198/1530 train_loss:4.1744 train_time:30674ms step_avg:163.16ms
step:199/1530 train_loss:4.0263 train_time:30837ms step_avg:163.16ms
step:200/1530 train_loss:4.0997 train_time:31000ms step_avg:163.16ms
step:201/1530 train_loss:4.0174 train_time:31163ms step_avg:163.16ms
step:202/1530 train_loss:4.2564 train_time:31326ms step_avg:163.16ms
step:203/1530 train_loss:4.0725 train_time:31488ms step_avg:163.15ms
step:204/1530 train_loss:4.1876 train_time:31651ms step_avg:163.15ms
step:205/1530 train_loss:4.2418 train_time:31814ms step_avg:163.15ms
step:206/1530 train_loss:3.9348 train_time:31978ms step_avg:163.15ms
step:207/1530 train_loss:4.0736 train_time:32140ms step_avg:163.15ms
step:208/1530 train_loss:4.0902 train_time:32303ms step_avg:163.15ms
step:209/1530 train_loss:4.2218 train_time:32466ms step_avg:163.14ms
step:210/1530 train_loss:4.1591 train_time:32632ms step_avg:163.16ms
step:211/1530 train_loss:4.0484 train_time:32795ms step_avg:163.16ms
step:212/1530 train_loss:4.1044 train_time:32959ms step_avg:163.16ms
step:213/1530 train_loss:4.0415 train_time:33120ms step_avg:163.15ms
step:214/1530 train_loss:4.0994 train_time:33283ms step_avg:163.15ms
step:215/1530 train_loss:3.9445 train_time:33447ms step_avg:163.16ms
step:216/1530 train_loss:3.9900 train_time:33611ms step_avg:163.16ms
step:217/1530 train_loss:3.9996 train_time:33775ms step_avg:163.16ms
step:218/1530 train_loss:4.0777 train_time:33937ms step_avg:163.16ms
step:219/1530 train_loss:4.0678 train_time:34100ms step_avg:163.16ms
step:220/1530 train_loss:4.0736 train_time:34263ms step_avg:163.16ms
step:221/1530 train_loss:4.0743 train_time:34426ms step_avg:163.16ms
step:222/1530 train_loss:3.9933 train_time:34589ms step_avg:163.15ms
step:223/1530 train_loss:3.9884 train_time:34752ms step_avg:163.16ms
step:224/1530 train_loss:4.3055 train_time:34915ms step_avg:163.15ms
step:225/1530 train_loss:3.9040 train_time:35078ms step_avg:163.15ms
step:226/1530 train_loss:3.9822 train_time:35240ms step_avg:163.15ms
step:227/1530 train_loss:3.9711 train_time:35403ms step_avg:163.15ms
step:228/1530 train_loss:4.1305 train_time:35568ms step_avg:163.16ms
step:229/1530 train_loss:3.9106 train_time:35735ms step_avg:163.17ms
step:230/1530 train_loss:4.0281 train_time:35900ms step_avg:163.18ms
step:231/1530 train_loss:3.8969 train_time:36067ms step_avg:163.20ms
step:232/1530 train_loss:3.9533 train_time:36233ms step_avg:163.21ms
step:233/1530 train_loss:4.0756 train_time:36399ms step_avg:163.22ms
step:234/1530 train_loss:4.0209 train_time:36566ms step_avg:163.24ms
step:235/1530 train_loss:3.8950 train_time:36734ms step_avg:163.26ms
step:236/1530 train_loss:4.0695 train_time:36900ms step_avg:163.27ms
step:237/1530 train_loss:4.0686 train_time:37066ms step_avg:163.28ms
step:238/1530 train_loss:3.9296 train_time:37231ms step_avg:163.30ms
step:239/1530 train_loss:4.0704 train_time:37397ms step_avg:163.31ms
step:240/1530 train_loss:4.1019 train_time:37563ms step_avg:163.32ms
step:241/1530 train_loss:3.9510 train_time:37727ms step_avg:163.32ms
step:242/1530 train_loss:4.1383 train_time:37895ms step_avg:163.34ms
step:243/1530 train_loss:4.0032 train_time:38061ms step_avg:163.35ms
step:244/1530 train_loss:4.0655 train_time:38227ms step_avg:163.36ms
step:245/1530 train_loss:4.1349 train_time:38393ms step_avg:163.38ms
step:246/1530 train_loss:4.0515 train_time:38560ms step_avg:163.39ms
step:247/1530 train_loss:4.0058 train_time:38725ms step_avg:163.40ms
step:248/1530 train_loss:4.1079 train_time:38892ms step_avg:163.41ms
step:249/1530 train_loss:3.9198 train_time:39057ms step_avg:163.42ms
step:250/1530 train_loss:3.9615 train_time:39223ms step_avg:163.43ms
step:250/1530 val_loss:3.9950 train_time:39271ms step_avg:163.63ms
step:251/1530 train_loss:4.0586 train_time:39393ms step_avg:163.46ms
step:252/1530 train_loss:4.1557 train_time:39560ms step_avg:163.47ms
step:253/1530 train_loss:3.9232 train_time:39727ms step_avg:163.49ms
step:254/1530 train_loss:3.8740 train_time:39895ms step_avg:163.50ms
step:255/1530 train_loss:4.0754 train_time:40060ms step_avg:163.51ms
step:256/1530 train_loss:3.9697 train_time:40226ms step_avg:163.52ms
step:257/1530 train_loss:3.9812 train_time:40392ms step_avg:163.53ms
step:258/1530 train_loss:3.9700 train_time:40559ms step_avg:163.54ms
step:259/1530 train_loss:4.0217 train_time:40725ms step_avg:163.55ms
step:260/1530 train_loss:4.0521 train_time:40893ms step_avg:163.57ms
step:261/1530 train_loss:4.0152 train_time:41059ms step_avg:163.58ms
step:262/1530 train_loss:3.9804 train_time:41226ms step_avg:163.59ms
step:263/1530 train_loss:3.8853 train_time:41391ms step_avg:163.60ms
step:264/1530 train_loss:3.9795 train_time:41558ms step_avg:163.61ms
step:265/1530 train_loss:3.8622 train_time:41725ms step_avg:163.63ms
step:266/1530 train_loss:3.9052 train_time:41892ms step_avg:163.64ms
step:267/1530 train_loss:3.9262 train_time:42058ms step_avg:163.65ms
step:268/1530 train_loss:3.9524 train_time:42223ms step_avg:163.65ms
step:269/1530 train_loss:3.8522 train_time:42389ms step_avg:163.66ms
step:270/1530 train_loss:4.0982 train_time:42555ms step_avg:163.67ms
step:271/1530 train_loss:3.9567 train_time:42721ms step_avg:163.68ms
step:272/1530 train_loss:3.9208 train_time:42887ms step_avg:163.69ms
step:273/1530 train_loss:3.9381 train_time:43053ms step_avg:163.70ms
step:274/1530 train_loss:4.0324 train_time:43220ms step_avg:163.71ms
step:275/1530 train_loss:4.0570 train_time:43386ms step_avg:163.72ms
step:276/1530 train_loss:4.2339 train_time:43555ms step_avg:163.74ms
step:277/1530 train_loss:4.0349 train_time:43721ms step_avg:163.75ms
step:278/1530 train_loss:4.0816 train_time:43887ms step_avg:163.76ms
step:279/1530 train_loss:3.9950 train_time:44054ms step_avg:163.77ms
step:280/1530 train_loss:4.2033 train_time:44222ms step_avg:163.78ms
step:281/1530 train_loss:3.9647 train_time:44389ms step_avg:163.80ms
step:282/1530 train_loss:3.9431 train_time:44556ms step_avg:163.81ms
step:283/1530 train_loss:3.9028 train_time:44722ms step_avg:163.82ms
step:284/1530 train_loss:4.0379 train_time:44887ms step_avg:163.82ms
step:285/1530 train_loss:4.0478 train_time:45054ms step_avg:163.83ms
step:286/1530 train_loss:4.0801 train_time:45219ms step_avg:163.84ms
step:287/1530 train_loss:3.8915 train_time:45384ms step_avg:163.84ms
step:288/1530 train_loss:4.0008 train_time:45550ms step_avg:163.85ms
step:289/1530 train_loss:3.8682 train_time:45716ms step_avg:163.86ms
step:290/1530 train_loss:3.8540 train_time:45880ms step_avg:163.86ms
step:291/1530 train_loss:3.9029 train_time:46045ms step_avg:163.86ms
step:292/1530 train_loss:3.8521 train_time:46211ms step_avg:163.87ms
step:293/1530 train_loss:3.8854 train_time:46376ms step_avg:163.87ms
step:294/1530 train_loss:3.9223 train_time:46542ms step_avg:163.88ms
step:295/1530 train_loss:3.8253 train_time:46707ms step_avg:163.88ms
step:296/1530 train_loss:3.8466 train_time:46874ms step_avg:163.89ms
step:297/1530 train_loss:3.8569 train_time:47039ms step_avg:163.90ms
step:298/1530 train_loss:3.9587 train_time:47204ms step_avg:163.90ms
step:299/1530 train_loss:3.8155 train_time:47370ms step_avg:163.91ms
step:300/1530 train_loss:3.9552 train_time:47536ms step_avg:163.92ms
step:301/1530 train_loss:3.9520 train_time:47701ms step_avg:163.92ms
step:302/1530 train_loss:3.9240 train_time:47865ms step_avg:163.92ms
step:303/1530 train_loss:3.9719 train_time:48031ms step_avg:163.93ms
step:304/1530 train_loss:3.9589 train_time:48196ms step_avg:163.93ms
step:305/1530 train_loss:4.4418 train_time:48361ms step_avg:163.94ms
step:306/1530 train_loss:3.9301 train_time:48525ms step_avg:163.94ms
step:307/1530 train_loss:3.8333 train_time:48690ms step_avg:163.94ms
step:308/1530 train_loss:3.9778 train_time:48857ms step_avg:163.95ms
step:309/1530 train_loss:3.8678 train_time:49022ms step_avg:163.95ms
step:310/1530 train_loss:4.0770 train_time:49186ms step_avg:163.95ms
step:311/1530 train_loss:3.9259 train_time:49352ms step_avg:163.96ms
step:312/1530 train_loss:3.8595 train_time:49517ms step_avg:163.96ms
step:313/1530 train_loss:3.9223 train_time:49681ms step_avg:163.96ms
step:314/1530 train_loss:4.0448 train_time:49845ms step_avg:163.96ms
step:315/1530 train_loss:3.9372 train_time:50011ms step_avg:163.97ms
step:316/1530 train_loss:3.7878 train_time:50175ms step_avg:163.97ms
step:317/1530 train_loss:3.8718 train_time:50340ms step_avg:163.97ms
step:318/1530 train_loss:3.9115 train_time:50505ms step_avg:163.98ms
step:319/1530 train_loss:3.8898 train_time:50671ms step_avg:163.99ms
step:320/1530 train_loss:4.0031 train_time:50837ms step_avg:163.99ms
step:321/1530 train_loss:3.9457 train_time:51001ms step_avg:163.99ms
step:322/1530 train_loss:3.9231 train_time:51167ms step_avg:164.00ms
step:323/1530 train_loss:3.9988 train_time:51334ms step_avg:164.01ms
step:324/1530 train_loss:3.9346 train_time:51499ms step_avg:164.01ms
step:325/1530 train_loss:4.0028 train_time:51664ms step_avg:164.01ms
step:326/1530 train_loss:3.8827 train_time:51831ms step_avg:164.02ms
step:327/1530 train_loss:4.3915 train_time:51996ms step_avg:164.03ms
step:328/1530 train_loss:4.0623 train_time:52161ms step_avg:164.03ms
step:329/1530 train_loss:3.7877 train_time:52326ms step_avg:164.03ms
step:330/1530 train_loss:3.7353 train_time:52492ms step_avg:164.04ms
step:331/1530 train_loss:3.9689 train_time:52657ms step_avg:164.04ms
step:332/1530 train_loss:3.9023 train_time:52822ms step_avg:164.04ms
step:333/1530 train_loss:3.8780 train_time:52987ms step_avg:164.05ms
step:334/1530 train_loss:3.8323 train_time:53152ms step_avg:164.05ms
step:335/1530 train_loss:3.9992 train_time:53319ms step_avg:164.06ms
step:336/1530 train_loss:3.9558 train_time:53483ms step_avg:164.06ms
step:337/1530 train_loss:4.4129 train_time:53650ms step_avg:164.07ms
step:338/1530 train_loss:3.9253 train_time:53815ms step_avg:164.07ms
step:339/1530 train_loss:3.8623 train_time:53979ms step_avg:164.07ms
step:340/1530 train_loss:3.9295 train_time:54144ms step_avg:164.07ms
step:341/1530 train_loss:3.8454 train_time:54312ms step_avg:164.08ms
step:342/1530 train_loss:3.8003 train_time:54479ms step_avg:164.09ms
step:343/1530 train_loss:3.8261 train_time:54648ms step_avg:164.11ms
step:344/1530 train_loss:3.9919 train_time:54817ms step_avg:164.12ms
step:345/1530 train_loss:3.8058 train_time:54986ms step_avg:164.14ms
step:346/1530 train_loss:3.7616 train_time:55154ms step_avg:164.15ms
step:347/1530 train_loss:3.7925 train_time:55323ms step_avg:164.16ms
step:348/1530 train_loss:3.8567 train_time:55491ms step_avg:164.17ms
step:349/1530 train_loss:3.8197 train_time:55659ms step_avg:164.18ms
step:350/1530 train_loss:3.5712 train_time:55827ms step_avg:164.20ms
step:351/1530 train_loss:3.8187 train_time:55996ms step_avg:164.21ms
step:352/1530 train_loss:4.1862 train_time:56163ms step_avg:164.22ms
step:353/1530 train_loss:3.6484 train_time:56332ms step_avg:164.23ms
step:354/1530 train_loss:3.9214 train_time:56498ms step_avg:164.24ms
step:355/1530 train_loss:3.7774 train_time:56667ms step_avg:164.25ms
step:356/1530 train_loss:3.8797 train_time:56836ms step_avg:164.27ms
step:357/1530 train_loss:3.7532 train_time:57003ms step_avg:164.27ms
step:358/1530 train_loss:3.8581 train_time:57171ms step_avg:164.29ms
step:359/1530 train_loss:3.7581 train_time:57341ms step_avg:164.30ms
step:360/1530 train_loss:3.4161 train_time:57511ms step_avg:164.32ms
step:361/1530 train_loss:4.0071 train_time:57679ms step_avg:164.33ms
step:362/1530 train_loss:3.9072 train_time:57846ms step_avg:164.34ms
step:363/1530 train_loss:3.8297 train_time:58014ms step_avg:164.35ms
step:364/1530 train_loss:3.7332 train_time:58181ms step_avg:164.35ms
step:365/1530 train_loss:3.9056 train_time:58349ms step_avg:164.36ms
step:366/1530 train_loss:3.8513 train_time:58517ms step_avg:164.37ms
step:367/1530 train_loss:3.8445 train_time:58684ms step_avg:164.38ms
step:368/1530 train_loss:3.8376 train_time:58853ms step_avg:164.39ms
step:369/1530 train_loss:3.7360 train_time:59022ms step_avg:164.41ms
step:370/1530 train_loss:3.8707 train_time:59190ms step_avg:164.42ms
step:371/1530 train_loss:3.7230 train_time:59358ms step_avg:164.43ms
step:372/1530 train_loss:3.6862 train_time:59526ms step_avg:164.44ms
step:373/1530 train_loss:3.9083 train_time:59695ms step_avg:164.45ms
step:374/1530 train_loss:3.8236 train_time:59863ms step_avg:164.46ms
step:375/1530 train_loss:3.7935 train_time:60031ms step_avg:164.47ms
step:375/1530 val_loss:3.8147 train_time:60080ms step_avg:164.60ms
step:376/1530 train_loss:3.8593 train_time:60202ms step_avg:164.49ms
step:377/1530 train_loss:3.7817 train_time:60504ms step_avg:164.86ms
step:378/1530 train_loss:3.8391 train_time:60682ms step_avg:164.90ms
step:379/1530 train_loss:3.8662 train_time:60999ms step_avg:165.31ms
step:380/1530 train_loss:3.9494 train_time:61165ms step_avg:165.31ms
step:381/1530 train_loss:3.8335 train_time:61333ms step_avg:165.32ms
step:382/1530 train_loss:3.8000 train_time:61504ms step_avg:165.33ms
step:383/1530 train_loss:3.7887 train_time:61671ms step_avg:165.34ms
step:384/1530 train_loss:3.8613 train_time:61838ms step_avg:165.34ms
step:385/1530 train_loss:3.7894 train_time:62007ms step_avg:165.35ms
step:386/1530 train_loss:3.8843 train_time:62173ms step_avg:165.35ms
step:387/1530 train_loss:4.0467 train_time:62342ms step_avg:165.36ms
step:388/1530 train_loss:3.7856 train_time:62510ms step_avg:165.37ms
step:389/1530 train_loss:3.7866 train_time:62679ms step_avg:165.38ms
step:390/1530 train_loss:3.8898 train_time:62847ms step_avg:165.39ms
step:391/1530 train_loss:3.8051 train_time:63014ms step_avg:165.39ms
step:392/1530 train_loss:3.9200 train_time:63182ms step_avg:165.40ms
step:393/1530 train_loss:3.7606 train_time:63349ms step_avg:165.40ms
step:394/1530 train_loss:3.8742 train_time:63515ms step_avg:165.40ms
step:395/1530 train_loss:3.6293 train_time:63684ms step_avg:165.41ms
step:396/1530 train_loss:3.8288 train_time:63852ms step_avg:165.42ms
step:397/1530 train_loss:3.8527 train_time:64020ms step_avg:165.43ms
step:398/1530 train_loss:3.8673 train_time:64187ms step_avg:165.43ms
step:399/1530 train_loss:3.7571 train_time:64353ms step_avg:165.43ms
step:400/1530 train_loss:3.8186 train_time:64522ms step_avg:165.44ms
step:401/1530 train_loss:3.8976 train_time:64689ms step_avg:165.45ms
step:402/1530 train_loss:3.8317 train_time:64857ms step_avg:165.45ms
step:403/1530 train_loss:3.9501 train_time:65024ms step_avg:165.46ms
step:404/1530 train_loss:3.6668 train_time:65190ms step_avg:165.46ms
step:405/1530 train_loss:3.7744 train_time:65360ms step_avg:165.47ms
step:406/1530 train_loss:4.0772 train_time:65527ms step_avg:165.47ms
step:407/1530 train_loss:3.7662 train_time:65694ms step_avg:165.48ms
step:408/1530 train_loss:3.8135 train_time:65861ms step_avg:165.48ms
step:409/1530 train_loss:3.8491 train_time:66026ms step_avg:165.48ms
step:410/1530 train_loss:3.7491 train_time:66193ms step_avg:165.48ms
step:411/1530 train_loss:3.7544 train_time:66361ms step_avg:165.49ms
step:412/1530 train_loss:4.1728 train_time:66528ms step_avg:165.49ms
step:413/1530 train_loss:3.6093 train_time:66695ms step_avg:165.50ms
step:414/1530 train_loss:4.0009 train_time:66863ms step_avg:165.50ms
step:415/1530 train_loss:3.7506 train_time:67029ms step_avg:165.50ms
step:416/1530 train_loss:3.7558 train_time:67196ms step_avg:165.51ms
step:417/1530 train_loss:3.9486 train_time:67364ms step_avg:165.51ms
step:418/1530 train_loss:3.6799 train_time:67530ms step_avg:165.52ms
step:419/1530 train_loss:3.7970 train_time:67698ms step_avg:165.52ms
step:420/1530 train_loss:3.6894 train_time:67865ms step_avg:165.52ms
step:421/1530 train_loss:3.6410 train_time:68032ms step_avg:165.53ms
step:422/1530 train_loss:3.7717 train_time:68200ms step_avg:165.53ms
step:423/1530 train_loss:3.8643 train_time:68367ms step_avg:165.54ms
step:424/1530 train_loss:3.6038 train_time:68535ms step_avg:165.54ms
step:425/1530 train_loss:3.7908 train_time:68703ms step_avg:165.55ms
step:426/1530 train_loss:3.6344 train_time:68870ms step_avg:165.55ms
step:427/1530 train_loss:3.8798 train_time:69038ms step_avg:165.56ms
step:428/1530 train_loss:3.7991 train_time:69205ms step_avg:165.56ms
step:429/1530 train_loss:3.7456 train_time:69372ms step_avg:165.57ms
step:430/1530 train_loss:3.6943 train_time:69540ms step_avg:165.57ms
step:431/1530 train_loss:3.6164 train_time:69707ms step_avg:165.58ms
step:432/1530 train_loss:3.7544 train_time:69875ms step_avg:165.58ms
step:433/1530 train_loss:3.8074 train_time:70042ms step_avg:165.58ms
step:434/1530 train_loss:3.7687 train_time:70209ms step_avg:165.59ms
step:435/1530 train_loss:3.7974 train_time:70376ms step_avg:165.59ms
step:436/1530 train_loss:3.8232 train_time:70544ms step_avg:165.60ms
step:437/1530 train_loss:3.7113 train_time:70710ms step_avg:165.60ms
step:438/1530 train_loss:3.6902 train_time:70878ms step_avg:165.60ms
step:439/1530 train_loss:3.7093 train_time:71045ms step_avg:165.61ms
step:440/1530 train_loss:3.8847 train_time:71212ms step_avg:165.61ms
step:441/1530 train_loss:3.7519 train_time:71382ms step_avg:165.62ms
step:442/1530 train_loss:3.7328 train_time:71548ms step_avg:165.62ms
step:443/1530 train_loss:3.6118 train_time:71715ms step_avg:165.62ms
step:444/1530 train_loss:3.9141 train_time:71882ms step_avg:165.63ms
step:445/1530 train_loss:3.8353 train_time:72048ms step_avg:165.63ms
step:446/1530 train_loss:3.8207 train_time:72216ms step_avg:165.63ms
step:447/1530 train_loss:3.7399 train_time:72384ms step_avg:165.64ms
step:448/1530 train_loss:3.8437 train_time:72550ms step_avg:165.64ms
step:449/1530 train_loss:3.6842 train_time:72720ms step_avg:165.65ms
step:450/1530 train_loss:3.7028 train_time:72886ms step_avg:165.65ms
step:451/1530 train_loss:3.5767 train_time:73054ms step_avg:165.66ms
step:452/1530 train_loss:3.6984 train_time:73223ms step_avg:165.66ms
step:453/1530 train_loss:3.6676 train_time:73390ms step_avg:165.67ms
step:454/1530 train_loss:3.6340 train_time:73558ms step_avg:165.67ms
step:455/1530 train_loss:3.8317 train_time:73726ms step_avg:165.68ms
step:456/1530 train_loss:3.7185 train_time:73896ms step_avg:165.69ms
step:457/1530 train_loss:3.7718 train_time:74066ms step_avg:165.70ms
step:458/1530 train_loss:3.8188 train_time:74235ms step_avg:165.70ms
step:459/1530 train_loss:3.6240 train_time:74406ms step_avg:165.71ms
step:460/1530 train_loss:3.7795 train_time:74575ms step_avg:165.72ms
step:461/1530 train_loss:3.6801 train_time:74745ms step_avg:165.73ms
step:462/1530 train_loss:3.7309 train_time:74914ms step_avg:165.74ms
step:463/1530 train_loss:3.7660 train_time:75086ms step_avg:165.75ms
step:464/1530 train_loss:3.7092 train_time:75255ms step_avg:165.76ms
step:465/1530 train_loss:3.7058 train_time:75424ms step_avg:165.77ms
step:466/1530 train_loss:3.7866 train_time:75594ms step_avg:165.78ms
step:467/1530 train_loss:3.8182 train_time:75765ms step_avg:165.79ms
step:468/1530 train_loss:3.7781 train_time:75934ms step_avg:165.80ms
step:469/1530 train_loss:3.6749 train_time:76104ms step_avg:165.80ms
step:470/1530 train_loss:3.7502 train_time:76275ms step_avg:165.81ms
step:471/1530 train_loss:3.8008 train_time:76444ms step_avg:165.82ms
step:472/1530 train_loss:3.7772 train_time:76615ms step_avg:165.83ms
step:473/1530 train_loss:3.7056 train_time:76785ms step_avg:165.84ms
step:474/1530 train_loss:3.5830 train_time:76953ms step_avg:165.85ms
step:475/1530 train_loss:4.0062 train_time:77124ms step_avg:165.86ms
step:476/1530 train_loss:3.7443 train_time:77293ms step_avg:165.87ms
step:477/1530 train_loss:3.5851 train_time:77465ms step_avg:165.88ms
step:478/1530 train_loss:3.8131 train_time:77634ms step_avg:165.89ms
step:479/1530 train_loss:3.7607 train_time:77804ms step_avg:165.89ms
step:480/1530 train_loss:3.9125 train_time:77974ms step_avg:165.90ms
step:481/1530 train_loss:3.7184 train_time:78143ms step_avg:165.91ms
step:482/1530 train_loss:3.5215 train_time:78312ms step_avg:165.92ms
step:483/1530 train_loss:3.7958 train_time:78482ms step_avg:165.92ms
step:484/1530 train_loss:3.6539 train_time:78651ms step_avg:165.93ms
step:485/1530 train_loss:3.6505 train_time:78822ms step_avg:165.94ms
step:486/1530 train_loss:3.5707 train_time:78991ms step_avg:165.95ms
step:487/1530 train_loss:3.6819 train_time:79162ms step_avg:165.96ms
step:488/1530 train_loss:3.8745 train_time:79330ms step_avg:165.96ms
step:489/1530 train_loss:3.7015 train_time:79500ms step_avg:165.97ms
step:490/1530 train_loss:3.5810 train_time:79669ms step_avg:165.98ms
step:491/1530 train_loss:3.6084 train_time:79838ms step_avg:165.98ms
step:492/1530 train_loss:3.7200 train_time:80009ms step_avg:165.99ms
step:493/1530 train_loss:3.5678 train_time:80181ms step_avg:166.01ms
step:494/1530 train_loss:3.6993 train_time:80350ms step_avg:166.01ms
step:495/1530 train_loss:3.6478 train_time:80521ms step_avg:166.02ms
step:496/1530 train_loss:3.4979 train_time:80691ms step_avg:166.03ms
step:497/1530 train_loss:3.7299 train_time:80861ms step_avg:166.04ms
step:498/1530 train_loss:3.7789 train_time:81030ms step_avg:166.04ms
step:499/1530 train_loss:3.8108 train_time:81201ms step_avg:166.06ms
step:500/1530 train_loss:3.7308 train_time:81373ms step_avg:166.07ms
step:500/1530 val_loss:3.6956 train_time:81421ms step_avg:166.17ms
step:501/1530 train_loss:3.7960 train_time:81543ms step_avg:166.08ms
step:502/1530 train_loss:3.7385 train_time:81713ms step_avg:166.08ms
step:503/1530 train_loss:3.7694 train_time:81884ms step_avg:166.09ms
step:504/1530 train_loss:3.7096 train_time:82053ms step_avg:166.10ms
step:505/1530 train_loss:3.7971 train_time:82223ms step_avg:166.11ms
step:506/1530 train_loss:3.6303 train_time:82393ms step_avg:166.11ms
step:507/1530 train_loss:3.7562 train_time:82561ms step_avg:166.12ms
step:508/1530 train_loss:3.8156 train_time:82732ms step_avg:166.13ms
step:509/1530 train_loss:3.7626 train_time:82901ms step_avg:166.13ms
step:510/1530 train_loss:3.5705 train_time:83071ms step_avg:166.14ms
step:511/1530 train_loss:3.7673 train_time:83240ms step_avg:166.15ms
step:512/1530 train_loss:3.7094 train_time:83410ms step_avg:166.16ms
step:513/1530 train_loss:3.6569 train_time:83579ms step_avg:166.16ms
step:514/1530 train_loss:3.8460 train_time:83750ms step_avg:166.17ms
step:515/1530 train_loss:3.7290 train_time:83920ms step_avg:166.18ms
step:516/1530 train_loss:4.0656 train_time:84089ms step_avg:166.18ms
step:517/1530 train_loss:3.6806 train_time:84258ms step_avg:166.19ms
step:518/1530 train_loss:3.7612 train_time:84427ms step_avg:166.19ms
step:519/1530 train_loss:3.6406 train_time:84596ms step_avg:166.20ms
step:520/1530 train_loss:3.6752 train_time:84765ms step_avg:166.21ms
step:521/1530 train_loss:3.6552 train_time:84934ms step_avg:166.21ms
step:522/1530 train_loss:3.6521 train_time:85105ms step_avg:166.22ms
step:523/1530 train_loss:4.2871 train_time:85276ms step_avg:166.23ms
step:524/1530 train_loss:3.7267 train_time:85446ms step_avg:166.24ms
step:525/1530 train_loss:3.6702 train_time:85614ms step_avg:166.24ms
step:526/1530 train_loss:3.6902 train_time:85783ms step_avg:166.25ms
step:527/1530 train_loss:3.6433 train_time:85952ms step_avg:166.25ms
step:528/1530 train_loss:3.6192 train_time:86121ms step_avg:166.26ms
step:529/1530 train_loss:3.8468 train_time:86290ms step_avg:166.26ms
step:530/1530 train_loss:3.6423 train_time:86461ms step_avg:166.27ms
step:531/1530 train_loss:3.9113 train_time:86631ms step_avg:166.28ms
step:532/1530 train_loss:3.7188 train_time:86800ms step_avg:166.28ms
step:533/1530 train_loss:3.6481 train_time:86968ms step_avg:166.29ms
step:534/1530 train_loss:3.6640 train_time:87136ms step_avg:166.29ms
step:535/1530 train_loss:3.6018 train_time:87306ms step_avg:166.30ms
step:536/1530 train_loss:3.7409 train_time:87477ms step_avg:166.31ms
step:537/1530 train_loss:3.7128 train_time:87646ms step_avg:166.31ms
step:538/1530 train_loss:3.6122 train_time:87819ms step_avg:166.32ms
step:539/1530 train_loss:4.1077 train_time:87989ms step_avg:166.33ms
step:540/1530 train_loss:3.6665 train_time:88157ms step_avg:166.33ms
step:541/1530 train_loss:3.7735 train_time:88327ms step_avg:166.34ms
step:542/1530 train_loss:3.5778 train_time:88497ms step_avg:166.35ms
step:543/1530 train_loss:3.5745 train_time:88664ms step_avg:166.35ms
step:544/1530 train_loss:3.6253 train_time:88834ms step_avg:166.36ms
step:545/1530 train_loss:3.5865 train_time:89003ms step_avg:166.36ms
step:546/1530 train_loss:3.6128 train_time:89172ms step_avg:166.37ms
step:547/1530 train_loss:3.6279 train_time:89342ms step_avg:166.37ms
step:548/1530 train_loss:3.5945 train_time:89511ms step_avg:166.38ms
step:549/1530 train_loss:3.7133 train_time:89680ms step_avg:166.38ms
step:550/1530 train_loss:3.6136 train_time:89849ms step_avg:166.39ms
step:551/1530 train_loss:3.6213 train_time:90020ms step_avg:166.39ms
step:552/1530 train_loss:3.9247 train_time:90188ms step_avg:166.40ms
step:553/1530 train_loss:3.7454 train_time:90357ms step_avg:166.40ms
step:554/1530 train_loss:3.7027 train_time:90526ms step_avg:166.41ms
step:555/1530 train_loss:3.6141 train_time:90694ms step_avg:166.41ms
step:556/1530 train_loss:3.6871 train_time:90863ms step_avg:166.42ms
step:557/1530 train_loss:3.3049 train_time:91032ms step_avg:166.42ms
step:558/1530 train_loss:3.6040 train_time:91202ms step_avg:166.43ms
step:559/1530 train_loss:3.6413 train_time:91370ms step_avg:166.43ms
step:560/1530 train_loss:3.6816 train_time:91540ms step_avg:166.44ms
step:561/1530 train_loss:3.6050 train_time:91709ms step_avg:166.44ms
step:562/1530 train_loss:3.5447 train_time:91878ms step_avg:166.45ms
step:563/1530 train_loss:3.7489 train_time:92047ms step_avg:166.45ms
step:564/1530 train_loss:3.5617 train_time:92217ms step_avg:166.46ms
step:565/1530 train_loss:3.6724 train_time:92385ms step_avg:166.46ms
step:566/1530 train_loss:3.6134 train_time:92685ms step_avg:166.70ms
step:567/1530 train_loss:3.5912 train_time:92863ms step_avg:166.72ms
step:568/1530 train_loss:3.6812 train_time:93034ms step_avg:166.73ms
step:569/1530 train_loss:3.6439 train_time:93355ms step_avg:167.00ms
step:570/1530 train_loss:3.6864 train_time:93525ms step_avg:167.01ms
step:571/1530 train_loss:3.7519 train_time:93695ms step_avg:167.01ms
step:572/1530 train_loss:3.7201 train_time:93865ms step_avg:167.02ms
step:573/1530 train_loss:3.7336 train_time:94036ms step_avg:167.03ms
step:574/1530 train_loss:3.7704 train_time:94209ms step_avg:167.04ms
step:575/1530 train_loss:3.7217 train_time:94383ms step_avg:167.05ms
step:576/1530 train_loss:3.7547 train_time:94554ms step_avg:167.06ms
step:577/1530 train_loss:3.6580 train_time:94726ms step_avg:167.06ms
step:578/1530 train_loss:3.6662 train_time:94897ms step_avg:167.07ms
step:579/1530 train_loss:3.6631 train_time:95067ms step_avg:167.08ms
step:580/1530 train_loss:3.5797 train_time:95238ms step_avg:167.08ms
step:581/1530 train_loss:3.6333 train_time:95409ms step_avg:167.09ms
step:582/1530 train_loss:3.8373 train_time:95581ms step_avg:167.10ms
step:583/1530 train_loss:3.6193 train_time:95753ms step_avg:167.11ms
step:584/1530 train_loss:3.5782 train_time:95925ms step_avg:167.12ms
step:585/1530 train_loss:3.7800 train_time:96096ms step_avg:167.12ms
step:586/1530 train_loss:3.5109 train_time:96267ms step_avg:167.13ms
step:587/1530 train_loss:3.6638 train_time:96439ms step_avg:167.14ms
step:588/1530 train_loss:3.6316 train_time:96609ms step_avg:167.14ms
step:589/1530 train_loss:3.9870 train_time:96782ms step_avg:167.15ms
step:590/1530 train_loss:3.7729 train_time:96954ms step_avg:167.16ms
step:591/1530 train_loss:3.4990 train_time:97125ms step_avg:167.17ms
step:592/1530 train_loss:3.5222 train_time:97299ms step_avg:167.18ms
step:593/1530 train_loss:3.4954 train_time:97470ms step_avg:167.19ms
step:594/1530 train_loss:3.5432 train_time:97642ms step_avg:167.20ms
step:595/1530 train_loss:3.9116 train_time:97814ms step_avg:167.20ms
step:596/1530 train_loss:3.6382 train_time:97986ms step_avg:167.21ms
step:597/1530 train_loss:3.5743 train_time:98156ms step_avg:167.22ms
step:598/1530 train_loss:3.6508 train_time:98326ms step_avg:167.22ms
step:599/1530 train_loss:3.4680 train_time:98499ms step_avg:167.23ms
step:600/1530 train_loss:3.5898 train_time:98669ms step_avg:167.24ms
step:601/1530 train_loss:3.6407 train_time:98843ms step_avg:167.25ms
step:602/1530 train_loss:3.6594 train_time:99016ms step_avg:167.26ms
step:603/1530 train_loss:3.7775 train_time:99186ms step_avg:167.26ms
step:604/1530 train_loss:3.6038 train_time:99358ms step_avg:167.27ms
step:605/1530 train_loss:3.6028 train_time:99531ms step_avg:167.28ms
step:606/1530 train_loss:3.5683 train_time:99705ms step_avg:167.29ms
step:607/1530 train_loss:3.8344 train_time:99876ms step_avg:167.30ms
step:608/1530 train_loss:3.6272 train_time:100047ms step_avg:167.30ms
step:609/1530 train_loss:3.6066 train_time:100220ms step_avg:167.31ms
step:610/1530 train_loss:3.6923 train_time:100389ms step_avg:167.31ms
step:611/1530 train_loss:3.5931 train_time:100561ms step_avg:167.32ms
step:612/1530 train_loss:3.5650 train_time:100731ms step_avg:167.33ms
step:613/1530 train_loss:3.7512 train_time:100902ms step_avg:167.33ms
step:614/1530 train_loss:3.6891 train_time:101073ms step_avg:167.34ms
step:615/1530 train_loss:3.6822 train_time:101244ms step_avg:167.35ms
step:616/1530 train_loss:3.6172 train_time:101414ms step_avg:167.35ms
step:617/1530 train_loss:3.5528 train_time:101585ms step_avg:167.36ms
step:618/1530 train_loss:3.6793 train_time:101755ms step_avg:167.36ms
step:619/1530 train_loss:3.5404 train_time:101926ms step_avg:167.37ms
step:620/1530 train_loss:3.5819 train_time:102098ms step_avg:167.37ms
step:621/1530 train_loss:3.9161 train_time:102269ms step_avg:167.38ms
step:622/1530 train_loss:3.5688 train_time:102442ms step_avg:167.39ms
step:623/1530 train_loss:3.5891 train_time:102615ms step_avg:167.40ms
step:624/1530 train_loss:3.6858 train_time:102785ms step_avg:167.40ms
step:625/1530 train_loss:3.6897 train_time:102956ms step_avg:167.41ms
step:625/1530 val_loss:3.6149 train_time:103005ms step_avg:167.49ms
step:626/1530 train_loss:3.7347 train_time:103128ms step_avg:167.42ms
step:627/1530 train_loss:3.7068 train_time:103301ms step_avg:167.42ms
step:628/1530 train_loss:3.7536 train_time:103470ms step_avg:167.43ms
step:629/1530 train_loss:3.5840 train_time:103642ms step_avg:167.44ms
step:630/1530 train_loss:3.7151 train_time:103812ms step_avg:167.44ms
step:631/1530 train_loss:3.7343 train_time:103983ms step_avg:167.44ms
step:632/1530 train_loss:3.6318 train_time:104156ms step_avg:167.45ms
step:633/1530 train_loss:3.6004 train_time:104327ms step_avg:167.46ms
step:634/1530 train_loss:3.6896 train_time:104497ms step_avg:167.46ms
step:635/1530 train_loss:3.9425 train_time:104667ms step_avg:167.47ms
step:636/1530 train_loss:3.5403 train_time:104837ms step_avg:167.47ms
step:637/1530 train_loss:3.3478 train_time:105009ms step_avg:167.48ms
step:638/1530 train_loss:3.5839 train_time:105179ms step_avg:167.48ms
step:639/1530 train_loss:3.6248 train_time:105349ms step_avg:167.49ms
step:640/1530 train_loss:3.5648 train_time:105519ms step_avg:167.49ms
step:641/1530 train_loss:3.5803 train_time:105687ms step_avg:167.49ms
step:642/1530 train_loss:3.6177 train_time:105857ms step_avg:167.50ms
step:643/1530 train_loss:3.5848 train_time:106028ms step_avg:167.50ms
step:644/1530 train_loss:3.5503 train_time:106197ms step_avg:167.50ms
step:645/1530 train_loss:3.7715 train_time:106369ms step_avg:167.51ms
step:646/1530 train_loss:3.6601 train_time:106540ms step_avg:167.52ms
step:647/1530 train_loss:3.6544 train_time:106710ms step_avg:167.52ms
step:648/1530 train_loss:3.7068 train_time:106883ms step_avg:167.53ms
step:649/1530 train_loss:3.7581 train_time:107053ms step_avg:167.53ms
step:650/1530 train_loss:3.6137 train_time:107224ms step_avg:167.54ms
step:651/1530 train_loss:3.7561 train_time:107395ms step_avg:167.54ms
step:652/1530 train_loss:3.5726 train_time:107565ms step_avg:167.55ms
step:653/1530 train_loss:3.6519 train_time:107734ms step_avg:167.55ms
step:654/1530 train_loss:3.4205 train_time:107906ms step_avg:167.56ms
step:655/1530 train_loss:3.5725 train_time:108075ms step_avg:167.56ms
step:656/1530 train_loss:3.5624 train_time:108247ms step_avg:167.56ms
step:657/1530 train_loss:3.4931 train_time:108418ms step_avg:167.57ms
step:658/1530 train_loss:3.6764 train_time:108588ms step_avg:167.57ms
step:659/1530 train_loss:3.5827 train_time:108759ms step_avg:167.58ms
step:660/1530 train_loss:3.6779 train_time:108931ms step_avg:167.59ms
step:661/1530 train_loss:3.7463 train_time:109102ms step_avg:167.59ms
step:662/1530 train_loss:3.6605 train_time:109272ms step_avg:167.60ms
step:663/1530 train_loss:3.5464 train_time:109442ms step_avg:167.60ms
step:664/1530 train_loss:3.5996 train_time:109612ms step_avg:167.60ms
step:665/1530 train_loss:3.4828 train_time:109784ms step_avg:167.61ms
step:666/1530 train_loss:3.7662 train_time:109953ms step_avg:167.61ms
step:667/1530 train_loss:3.5939 train_time:110125ms step_avg:167.62ms
step:668/1530 train_loss:3.6406 train_time:110295ms step_avg:167.62ms
step:669/1530 train_loss:3.4817 train_time:110468ms step_avg:167.63ms
step:670/1530 train_loss:3.5935 train_time:110638ms step_avg:167.63ms
step:671/1530 train_loss:3.5573 train_time:110810ms step_avg:167.64ms
step:672/1530 train_loss:3.5621 train_time:110982ms step_avg:167.65ms
step:673/1530 train_loss:3.8417 train_time:111152ms step_avg:167.65ms
step:674/1530 train_loss:3.6145 train_time:111322ms step_avg:167.65ms
step:675/1530 train_loss:3.6980 train_time:111492ms step_avg:167.66ms
step:676/1530 train_loss:3.4835 train_time:111662ms step_avg:167.66ms
step:677/1530 train_loss:3.5894 train_time:111834ms step_avg:167.67ms
step:678/1530 train_loss:3.5453 train_time:112005ms step_avg:167.67ms
step:679/1530 train_loss:3.6762 train_time:112176ms step_avg:167.68ms
step:680/1530 train_loss:3.5788 train_time:112346ms step_avg:167.68ms
step:681/1530 train_loss:3.6082 train_time:112518ms step_avg:167.69ms
step:682/1530 train_loss:3.6527 train_time:112691ms step_avg:167.70ms
step:683/1530 train_loss:3.7273 train_time:112864ms step_avg:167.70ms
step:684/1530 train_loss:3.6398 train_time:113035ms step_avg:167.71ms
step:685/1530 train_loss:3.6728 train_time:113210ms step_avg:167.72ms
step:686/1530 train_loss:3.6307 train_time:113382ms step_avg:167.72ms
step:687/1530 train_loss:3.6561 train_time:113553ms step_avg:167.73ms
step:688/1530 train_loss:3.2058 train_time:113731ms step_avg:167.74ms
step:689/1530 train_loss:3.3966 train_time:113905ms step_avg:167.75ms
step:690/1530 train_loss:3.5327 train_time:114079ms step_avg:167.76ms
step:691/1530 train_loss:3.4023 train_time:114251ms step_avg:167.77ms
step:692/1530 train_loss:3.6185 train_time:114423ms step_avg:167.78ms
step:693/1530 train_loss:3.6384 train_time:114595ms step_avg:167.78ms
step:694/1530 train_loss:3.5449 train_time:114767ms step_avg:167.79ms
step:695/1530 train_loss:3.5258 train_time:114937ms step_avg:167.79ms
step:696/1530 train_loss:3.8477 train_time:115111ms step_avg:167.80ms
step:697/1530 train_loss:3.5756 train_time:115284ms step_avg:167.81ms
step:698/1530 train_loss:3.6389 train_time:115455ms step_avg:167.81ms
step:699/1530 train_loss:3.7560 train_time:115631ms step_avg:167.82ms
step:700/1530 train_loss:3.5622 train_time:115803ms step_avg:167.83ms
step:701/1530 train_loss:3.5341 train_time:115974ms step_avg:167.84ms
step:702/1530 train_loss:3.5051 train_time:116147ms step_avg:167.84ms
step:703/1530 train_loss:3.4959 train_time:116319ms step_avg:167.85ms
step:704/1530 train_loss:3.5657 train_time:116493ms step_avg:167.86ms
step:705/1530 train_loss:3.5531 train_time:116669ms step_avg:167.87ms
step:706/1530 train_loss:3.5728 train_time:116844ms step_avg:167.88ms
step:707/1530 train_loss:3.6334 train_time:117019ms step_avg:167.89ms
step:708/1530 train_loss:3.5951 train_time:117192ms step_avg:167.90ms
step:709/1530 train_loss:3.5735 train_time:117367ms step_avg:167.91ms
step:710/1530 train_loss:3.5338 train_time:117538ms step_avg:167.91ms
step:711/1530 train_loss:3.5846 train_time:117712ms step_avg:167.92ms
step:712/1530 train_loss:3.6393 train_time:117888ms step_avg:167.93ms
step:713/1530 train_loss:3.6442 train_time:118066ms step_avg:167.95ms
step:714/1530 train_loss:3.5538 train_time:118237ms step_avg:167.95ms
step:715/1530 train_loss:3.5640 train_time:118410ms step_avg:167.96ms
step:716/1530 train_loss:3.5748 train_time:118582ms step_avg:167.96ms
step:717/1530 train_loss:3.6944 train_time:118756ms step_avg:167.97ms
step:718/1530 train_loss:3.5891 train_time:118928ms step_avg:167.98ms
step:719/1530 train_loss:3.6726 train_time:119101ms step_avg:167.98ms
step:720/1530 train_loss:3.8459 train_time:119275ms step_avg:167.99ms
step:721/1530 train_loss:3.4567 train_time:119449ms step_avg:168.00ms
step:722/1530 train_loss:3.7249 train_time:119621ms step_avg:168.01ms
step:723/1530 train_loss:3.7556 train_time:119792ms step_avg:168.01ms
step:724/1530 train_loss:3.5622 train_time:119966ms step_avg:168.02ms
step:725/1530 train_loss:3.6466 train_time:120138ms step_avg:168.03ms
step:726/1530 train_loss:3.5235 train_time:120312ms step_avg:168.03ms
step:727/1530 train_loss:3.5737 train_time:120488ms step_avg:168.04ms
step:728/1530 train_loss:3.7178 train_time:120661ms step_avg:168.05ms
step:729/1530 train_loss:3.6592 train_time:120833ms step_avg:168.06ms
step:730/1530 train_loss:3.6528 train_time:121008ms step_avg:168.07ms
step:731/1530 train_loss:3.5491 train_time:121180ms step_avg:168.07ms
step:732/1530 train_loss:3.5887 train_time:121353ms step_avg:168.08ms
step:733/1530 train_loss:3.8163 train_time:121529ms step_avg:168.09ms
step:734/1530 train_loss:3.5550 train_time:121704ms step_avg:168.10ms
step:735/1530 train_loss:3.6068 train_time:121875ms step_avg:168.10ms
step:736/1530 train_loss:3.7343 train_time:122049ms step_avg:168.11ms
step:737/1530 train_loss:3.6691 train_time:122221ms step_avg:168.12ms
step:738/1530 train_loss:3.5931 train_time:122391ms step_avg:168.12ms
step:739/1530 train_loss:3.4950 train_time:122564ms step_avg:168.13ms
step:740/1530 train_loss:4.1049 train_time:122741ms step_avg:168.14ms
step:741/1530 train_loss:3.4806 train_time:122914ms step_avg:168.14ms
step:742/1530 train_loss:3.5411 train_time:123088ms step_avg:168.15ms
step:743/1530 train_loss:3.5782 train_time:123261ms step_avg:168.16ms
step:744/1530 train_loss:3.6402 train_time:123434ms step_avg:168.17ms
step:745/1530 train_loss:3.5750 train_time:123609ms step_avg:168.18ms
step:746/1530 train_loss:3.5865 train_time:123781ms step_avg:168.18ms
step:747/1530 train_loss:3.6367 train_time:123956ms step_avg:168.19ms
step:748/1530 train_loss:3.5523 train_time:124134ms step_avg:168.20ms
step:749/1530 train_loss:3.5517 train_time:124307ms step_avg:168.21ms
step:750/1530 train_loss:3.5899 train_time:124479ms step_avg:168.21ms
step:750/1530 val_loss:3.5587 train_time:124528ms step_avg:168.28ms
step:751/1530 train_loss:3.5698 train_time:124653ms step_avg:168.22ms
step:752/1530 train_loss:3.6054 train_time:124824ms step_avg:168.23ms
step:753/1530 train_loss:3.6121 train_time:124996ms step_avg:168.23ms
step:754/1530 train_loss:3.5907 train_time:125170ms step_avg:168.24ms
step:755/1530 train_loss:3.6710 train_time:125473ms step_avg:168.42ms
step:756/1530 train_loss:3.4469 train_time:125658ms step_avg:168.44ms
step:757/1530 train_loss:3.7158 train_time:125831ms step_avg:168.45ms
step:758/1530 train_loss:3.6410 train_time:126001ms step_avg:168.45ms
step:759/1530 train_loss:3.5811 train_time:126324ms step_avg:168.66ms
step:760/1530 train_loss:3.7015 train_time:126494ms step_avg:168.66ms
step:761/1530 train_loss:3.3972 train_time:126667ms step_avg:168.67ms
step:762/1530 train_loss:3.5420 train_time:126838ms step_avg:168.67ms
step:763/1530 train_loss:3.6562 train_time:127011ms step_avg:168.67ms
step:764/1530 train_loss:3.3108 train_time:127183ms step_avg:168.68ms
step:765/1530 train_loss:3.7231 train_time:127356ms step_avg:168.68ms
step:766/1530 train_loss:3.5673 train_time:127531ms step_avg:168.69ms
step:767/1530 train_loss:3.5561 train_time:127703ms step_avg:168.70ms
step:768/1530 train_loss:3.5595 train_time:127878ms step_avg:168.70ms
step:769/1530 train_loss:3.5769 train_time:128052ms step_avg:168.71ms
step:770/1530 train_loss:3.6310 train_time:128224ms step_avg:168.72ms
step:771/1530 train_loss:3.8728 train_time:128397ms step_avg:168.72ms
step:772/1530 train_loss:3.4426 train_time:128570ms step_avg:168.73ms
step:773/1530 train_loss:3.6245 train_time:128741ms step_avg:168.73ms
step:774/1530 train_loss:3.6368 train_time:128913ms step_avg:168.73ms
step:775/1530 train_loss:3.6017 train_time:129084ms step_avg:168.74ms
step:776/1530 train_loss:3.3873 train_time:129257ms step_avg:168.74ms
step:777/1530 train_loss:3.3791 train_time:129432ms step_avg:168.75ms
step:778/1530 train_loss:3.4869 train_time:129603ms step_avg:168.75ms
step:779/1530 train_loss:3.5756 train_time:129776ms step_avg:168.76ms
step:780/1530 train_loss:3.5787 train_time:129950ms step_avg:168.77ms
step:781/1530 train_loss:3.6659 train_time:130120ms step_avg:168.77ms
step:782/1530 train_loss:3.5770 train_time:130294ms step_avg:168.77ms
step:783/1530 train_loss:3.5688 train_time:130467ms step_avg:168.78ms
step:784/1530 train_loss:3.6047 train_time:130637ms step_avg:168.78ms
step:785/1530 train_loss:3.5516 train_time:130809ms step_avg:168.79ms
step:786/1530 train_loss:3.4300 train_time:130983ms step_avg:168.79ms
step:787/1530 train_loss:3.7122 train_time:131155ms step_avg:168.80ms
step:788/1530 train_loss:3.4937 train_time:131329ms step_avg:168.80ms
step:789/1530 train_loss:3.5416 train_time:131501ms step_avg:168.81ms
step:790/1530 train_loss:3.6160 train_time:131674ms step_avg:168.81ms
step:791/1530 train_loss:3.7661 train_time:131851ms step_avg:168.82ms
step:792/1530 train_loss:3.7517 train_time:132023ms step_avg:168.83ms
step:793/1530 train_loss:3.4331 train_time:132194ms step_avg:168.83ms
step:794/1530 train_loss:3.5818 train_time:132367ms step_avg:168.84ms
step:795/1530 train_loss:3.6685 train_time:132541ms step_avg:168.84ms
step:796/1530 train_loss:3.7300 train_time:132716ms step_avg:168.85ms
step:797/1530 train_loss:3.5197 train_time:132891ms step_avg:168.86ms
step:798/1530 train_loss:3.6414 train_time:133067ms step_avg:168.87ms
step:799/1530 train_loss:3.5319 train_time:133242ms step_avg:168.87ms
step:800/1530 train_loss:3.5201 train_time:133414ms step_avg:168.88ms
step:801/1530 train_loss:3.6192 train_time:133589ms step_avg:168.89ms
step:802/1530 train_loss:3.4871 train_time:133764ms step_avg:168.89ms
step:803/1530 train_loss:3.4766 train_time:133937ms step_avg:168.90ms
step:804/1530 train_loss:3.6135 train_time:134112ms step_avg:168.91ms
step:805/1530 train_loss:3.5072 train_time:134288ms step_avg:168.92ms
step:806/1530 train_loss:3.5505 train_time:134459ms step_avg:168.92ms
step:807/1530 train_loss:3.6347 train_time:134634ms step_avg:168.93ms
step:808/1530 train_loss:3.5395 train_time:134811ms step_avg:168.94ms
step:809/1530 train_loss:3.4866 train_time:134984ms step_avg:168.94ms
step:810/1530 train_loss:3.5518 train_time:135156ms step_avg:168.95ms
step:811/1530 train_loss:3.5755 train_time:135331ms step_avg:168.95ms
step:812/1530 train_loss:3.5898 train_time:135505ms step_avg:168.96ms
step:813/1530 train_loss:3.6199 train_time:135677ms step_avg:168.96ms
step:814/1530 train_loss:3.5582 train_time:135853ms step_avg:168.97ms
step:815/1530 train_loss:3.5539 train_time:136026ms step_avg:168.98ms
step:816/1530 train_loss:3.6688 train_time:136200ms step_avg:168.98ms
step:817/1530 train_loss:3.7587 train_time:136374ms step_avg:168.99ms
step:818/1530 train_loss:3.5165 train_time:136546ms step_avg:168.99ms
step:819/1530 train_loss:3.7105 train_time:136720ms step_avg:169.00ms
step:820/1530 train_loss:3.4805 train_time:136897ms step_avg:169.01ms
step:821/1530 train_loss:3.5561 train_time:137071ms step_avg:169.01ms
step:822/1530 train_loss:3.6905 train_time:137247ms step_avg:169.02ms
step:823/1530 train_loss:3.5633 train_time:137419ms step_avg:169.03ms
step:824/1530 train_loss:3.5002 train_time:137592ms step_avg:169.03ms
step:825/1530 train_loss:3.6057 train_time:137768ms step_avg:169.04ms
step:826/1530 train_loss:3.4715 train_time:137944ms step_avg:169.05ms
step:827/1530 train_loss:3.7242 train_time:138118ms step_avg:169.05ms
step:828/1530 train_loss:3.6117 train_time:138293ms step_avg:169.06ms
step:829/1530 train_loss:3.6139 train_time:138470ms step_avg:169.07ms
step:830/1530 train_loss:3.5288 train_time:138645ms step_avg:169.08ms
step:831/1530 train_loss:3.5961 train_time:138818ms step_avg:169.08ms
step:832/1530 train_loss:3.5024 train_time:138994ms step_avg:169.09ms
step:833/1530 train_loss:3.6436 train_time:139171ms step_avg:169.10ms
step:834/1530 train_loss:3.4618 train_time:139343ms step_avg:169.11ms
step:835/1530 train_loss:3.4451 train_time:139517ms step_avg:169.11ms
step:836/1530 train_loss:3.7060 train_time:139693ms step_avg:169.12ms
step:837/1530 train_loss:3.3890 train_time:139869ms step_avg:169.13ms
step:838/1530 train_loss:3.5897 train_time:140041ms step_avg:169.13ms
step:839/1530 train_loss:3.4131 train_time:140214ms step_avg:169.14ms
step:840/1530 train_loss:3.4628 train_time:140387ms step_avg:169.14ms
step:841/1530 train_loss:3.5609 train_time:140559ms step_avg:169.14ms
step:842/1530 train_loss:3.5747 train_time:140735ms step_avg:169.15ms
step:843/1530 train_loss:3.5555 train_time:140906ms step_avg:169.16ms
step:844/1530 train_loss:3.4180 train_time:141080ms step_avg:169.16ms
step:845/1530 train_loss:3.6548 train_time:141254ms step_avg:169.17ms
step:846/1530 train_loss:3.5065 train_time:141430ms step_avg:169.17ms
step:847/1530 train_loss:3.4846 train_time:141605ms step_avg:169.18ms
step:848/1530 train_loss:3.6358 train_time:141779ms step_avg:169.19ms
step:849/1530 train_loss:3.4794 train_time:141954ms step_avg:169.19ms
step:850/1530 train_loss:3.4372 train_time:142131ms step_avg:169.20ms
step:851/1530 train_loss:3.7311 train_time:142302ms step_avg:169.21ms
step:852/1530 train_loss:3.4240 train_time:142477ms step_avg:169.21ms
step:853/1530 train_loss:3.5584 train_time:142649ms step_avg:169.22ms
step:854/1530 train_loss:3.6423 train_time:142826ms step_avg:169.22ms
step:855/1530 train_loss:3.5066 train_time:142998ms step_avg:169.23ms
step:856/1530 train_loss:3.5402 train_time:143172ms step_avg:169.23ms
step:857/1530 train_loss:3.5950 train_time:143346ms step_avg:169.24ms
step:858/1530 train_loss:3.4612 train_time:143522ms step_avg:169.25ms
step:859/1530 train_loss:3.5601 train_time:143695ms step_avg:169.25ms
step:860/1530 train_loss:3.5711 train_time:143869ms step_avg:169.26ms
step:861/1530 train_loss:3.6147 train_time:144046ms step_avg:169.27ms
step:862/1530 train_loss:3.5969 train_time:144223ms step_avg:169.28ms
step:863/1530 train_loss:3.5630 train_time:144399ms step_avg:169.28ms
step:864/1530 train_loss:3.3749 train_time:144574ms step_avg:169.29ms
step:865/1530 train_loss:3.5856 train_time:144745ms step_avg:169.29ms
step:866/1530 train_loss:3.9017 train_time:144922ms step_avg:169.30ms
step:867/1530 train_loss:3.4513 train_time:145095ms step_avg:169.31ms
step:868/1530 train_loss:3.6355 train_time:145269ms step_avg:169.31ms
step:869/1530 train_loss:3.6105 train_time:145441ms step_avg:169.31ms
step:870/1530 train_loss:3.4395 train_time:145617ms step_avg:169.32ms
step:871/1530 train_loss:3.3820 train_time:145792ms step_avg:169.33ms
step:872/1530 train_loss:3.6397 train_time:145968ms step_avg:169.34ms
step:873/1530 train_loss:3.4545 train_time:146139ms step_avg:169.34ms
step:874/1530 train_loss:3.2168 train_time:146315ms step_avg:169.35ms
step:875/1530 train_loss:3.6244 train_time:146489ms step_avg:169.35ms
step:875/1530 val_loss:3.5119 train_time:146540ms step_avg:169.41ms
step:876/1530 train_loss:3.4313 train_time:146664ms step_avg:169.36ms
step:877/1530 train_loss:3.6170 train_time:146838ms step_avg:169.36ms
step:878/1530 train_loss:3.4627 train_time:147013ms step_avg:169.37ms
step:879/1530 train_loss:3.6462 train_time:147185ms step_avg:169.37ms
step:880/1530 train_loss:3.2989 train_time:147358ms step_avg:169.38ms
step:881/1530 train_loss:3.4685 train_time:147529ms step_avg:169.38ms
step:882/1530 train_loss:3.6896 train_time:147701ms step_avg:169.38ms
step:883/1530 train_loss:3.8276 train_time:147875ms step_avg:169.39ms
step:884/1530 train_loss:3.5583 train_time:148050ms step_avg:169.39ms
step:885/1530 train_loss:3.4889 train_time:148222ms step_avg:169.40ms
step:886/1530 train_loss:3.5674 train_time:148397ms step_avg:169.40ms
step:887/1530 train_loss:4.0838 train_time:148570ms step_avg:169.41ms
step:888/1530 train_loss:3.8298 train_time:148749ms step_avg:169.42ms
step:889/1530 train_loss:3.5159 train_time:148922ms step_avg:169.42ms
step:890/1530 train_loss:3.5217 train_time:149094ms step_avg:169.43ms
step:891/1530 train_loss:3.3518 train_time:149267ms step_avg:169.43ms
step:892/1530 train_loss:3.7093 train_time:149440ms step_avg:169.43ms
step:893/1530 train_loss:3.4121 train_time:149612ms step_avg:169.44ms
step:894/1530 train_loss:3.6268 train_time:149789ms step_avg:169.44ms
step:895/1530 train_loss:3.6738 train_time:149962ms step_avg:169.45ms
step:896/1530 train_loss:3.4949 train_time:150135ms step_avg:169.45ms
step:897/1530 train_loss:3.5367 train_time:150310ms step_avg:169.46ms
step:898/1530 train_loss:3.5818 train_time:150486ms step_avg:169.47ms
step:899/1530 train_loss:3.4753 train_time:150659ms step_avg:169.47ms
step:900/1530 train_loss:3.4183 train_time:150830ms step_avg:169.47ms
step:901/1530 train_loss:3.6137 train_time:151003ms step_avg:169.48ms
step:902/1530 train_loss:3.6278 train_time:151176ms step_avg:169.48ms
step:903/1530 train_loss:3.5349 train_time:151351ms step_avg:169.49ms
step:904/1530 train_loss:3.4976 train_time:151523ms step_avg:169.49ms
step:905/1530 train_loss:3.4920 train_time:151695ms step_avg:169.49ms
step:906/1530 train_loss:3.6941 train_time:151870ms step_avg:169.50ms
step:907/1530 train_loss:3.5051 train_time:152042ms step_avg:169.50ms
step:908/1530 train_loss:3.5558 train_time:152215ms step_avg:169.50ms
step:909/1530 train_loss:3.4488 train_time:152390ms step_avg:169.51ms
step:910/1530 train_loss:3.5219 train_time:152570ms step_avg:169.52ms
step:911/1530 train_loss:3.6351 train_time:152745ms step_avg:169.53ms
step:912/1530 train_loss:3.5874 train_time:152922ms step_avg:169.54ms
step:913/1530 train_loss:3.4557 train_time:153101ms step_avg:169.55ms
step:914/1530 train_loss:3.7354 train_time:153280ms step_avg:169.56ms
step:915/1530 train_loss:3.5254 train_time:153459ms step_avg:169.57ms
step:916/1530 train_loss:3.6072 train_time:153635ms step_avg:169.57ms
step:917/1530 train_loss:3.5946 train_time:153808ms step_avg:169.58ms
step:918/1530 train_loss:4.8158 train_time:153987ms step_avg:169.59ms
step:919/1530 train_loss:3.4933 train_time:154166ms step_avg:169.60ms
step:920/1530 train_loss:3.5788 train_time:154340ms step_avg:169.60ms
step:921/1530 train_loss:3.5452 train_time:154517ms step_avg:169.61ms
step:922/1530 train_loss:3.5759 train_time:154693ms step_avg:169.62ms
step:923/1530 train_loss:3.6042 train_time:154870ms step_avg:169.63ms
step:924/1530 train_loss:3.6725 train_time:155045ms step_avg:169.63ms
step:925/1530 train_loss:3.6416 train_time:155219ms step_avg:169.64ms
step:926/1530 train_loss:3.5519 train_time:155393ms step_avg:169.64ms
step:927/1530 train_loss:3.5402 train_time:155567ms step_avg:169.65ms
step:928/1530 train_loss:3.7656 train_time:155744ms step_avg:169.66ms
step:929/1530 train_loss:3.6026 train_time:155920ms step_avg:169.66ms
step:930/1530 train_loss:3.3970 train_time:156098ms step_avg:169.67ms
step:931/1530 train_loss:3.4901 train_time:156271ms step_avg:169.68ms
step:932/1530 train_loss:3.6361 train_time:156449ms step_avg:169.68ms
step:933/1530 train_loss:3.3655 train_time:156624ms step_avg:169.69ms
step:934/1530 train_loss:3.5711 train_time:156801ms step_avg:169.70ms
step:935/1530 train_loss:3.4299 train_time:156979ms step_avg:169.71ms
step:936/1530 train_loss:3.5139 train_time:157157ms step_avg:169.72ms
step:937/1530 train_loss:3.6165 train_time:157336ms step_avg:169.73ms
step:938/1530 train_loss:3.5400 train_time:157511ms step_avg:169.73ms
step:939/1530 train_loss:3.6674 train_time:157690ms step_avg:169.74ms
step:940/1530 train_loss:3.4714 train_time:157865ms step_avg:169.75ms
step:941/1530 train_loss:3.5397 train_time:158041ms step_avg:169.75ms
step:942/1530 train_loss:3.3510 train_time:158217ms step_avg:169.76ms
step:943/1530 train_loss:3.7038 train_time:158398ms step_avg:169.77ms
step:944/1530 train_loss:3.3958 train_time:158709ms step_avg:169.92ms
step:945/1530 train_loss:3.4126 train_time:158890ms step_avg:169.94ms
step:946/1530 train_loss:5.0704 train_time:159069ms step_avg:169.95ms
step:947/1530 train_loss:3.5895 train_time:159245ms step_avg:169.95ms
step:948/1530 train_loss:3.4781 train_time:159420ms step_avg:169.96ms
step:949/1530 train_loss:3.3621 train_time:159743ms step_avg:170.12ms
step:950/1530 train_loss:3.4355 train_time:159918ms step_avg:170.13ms
step:951/1530 train_loss:3.3981 train_time:160097ms step_avg:170.13ms
step:952/1530 train_loss:3.4780 train_time:160273ms step_avg:170.14ms
step:953/1530 train_loss:3.5625 train_time:160448ms step_avg:170.15ms
step:954/1530 train_loss:3.4407 train_time:160627ms step_avg:170.16ms
step:955/1530 train_loss:3.4698 train_time:160802ms step_avg:170.16ms
step:956/1530 train_loss:3.4346 train_time:160978ms step_avg:170.17ms
step:957/1530 train_loss:3.4847 train_time:161158ms step_avg:170.18ms
step:958/1530 train_loss:3.4953 train_time:161337ms step_avg:170.19ms
step:959/1530 train_loss:3.5038 train_time:161515ms step_avg:170.20ms
step:960/1530 train_loss:3.3999 train_time:161691ms step_avg:170.20ms
step:961/1530 train_loss:3.6342 train_time:161865ms step_avg:170.21ms
step:962/1530 train_loss:3.5897 train_time:162041ms step_avg:170.21ms
step:963/1530 train_loss:3.6645 train_time:162219ms step_avg:170.22ms
step:964/1530 train_loss:3.4188 train_time:162398ms step_avg:170.23ms
step:965/1530 train_loss:3.4722 train_time:162571ms step_avg:170.23ms
step:966/1530 train_loss:3.6978 train_time:162745ms step_avg:170.24ms
step:967/1530 train_loss:3.5138 train_time:162921ms step_avg:170.24ms
step:968/1530 train_loss:3.5081 train_time:163096ms step_avg:170.25ms
step:969/1530 train_loss:3.5730 train_time:163270ms step_avg:170.25ms
step:970/1530 train_loss:3.3700 train_time:163443ms step_avg:170.25ms
step:971/1530 train_loss:3.5247 train_time:163617ms step_avg:170.26ms
step:972/1530 train_loss:3.4659 train_time:163790ms step_avg:170.26ms
step:973/1530 train_loss:3.5283 train_time:163963ms step_avg:170.26ms
step:974/1530 train_loss:3.5796 train_time:164139ms step_avg:170.27ms
step:975/1530 train_loss:3.4578 train_time:164315ms step_avg:170.27ms
step:976/1530 train_loss:3.6648 train_time:164488ms step_avg:170.28ms
step:977/1530 train_loss:3.5640 train_time:164662ms step_avg:170.28ms
step:978/1530 train_loss:3.3516 train_time:164837ms step_avg:170.29ms
step:979/1530 train_loss:3.6188 train_time:165013ms step_avg:170.29ms
step:980/1530 train_loss:3.4052 train_time:165188ms step_avg:170.30ms
step:981/1530 train_loss:3.5711 train_time:165364ms step_avg:170.30ms
step:982/1530 train_loss:3.5289 train_time:165539ms step_avg:170.31ms
step:983/1530 train_loss:3.5096 train_time:165716ms step_avg:170.31ms
step:984/1530 train_loss:3.4866 train_time:165891ms step_avg:170.32ms
step:985/1530 train_loss:3.5698 train_time:166066ms step_avg:170.32ms
step:986/1530 train_loss:3.4071 train_time:166242ms step_avg:170.33ms
step:987/1530 train_loss:3.4800 train_time:166414ms step_avg:170.33ms
step:988/1530 train_loss:3.4728 train_time:166588ms step_avg:170.34ms
step:989/1530 train_loss:3.4101 train_time:166761ms step_avg:170.34ms
step:990/1530 train_loss:3.6537 train_time:166938ms step_avg:170.34ms
step:991/1530 train_loss:3.4663 train_time:167113ms step_avg:170.35ms
step:992/1530 train_loss:3.4343 train_time:167291ms step_avg:170.36ms
step:993/1530 train_loss:3.4920 train_time:167470ms step_avg:170.37ms
step:994/1530 train_loss:3.5893 train_time:167645ms step_avg:170.37ms
step:995/1530 train_loss:3.5204 train_time:167818ms step_avg:170.37ms
step:996/1530 train_loss:3.4509 train_time:167990ms step_avg:170.38ms
step:997/1530 train_loss:3.7484 train_time:168165ms step_avg:170.38ms
step:998/1530 train_loss:3.4320 train_time:168338ms step_avg:170.38ms
step:999/1530 train_loss:3.5798 train_time:168515ms step_avg:170.39ms
step:1000/1530 train_loss:3.4326 train_time:168693ms step_avg:170.40ms
step:1000/1530 val_loss:3.4582 train_time:168743ms step_avg:170.45ms
step:1001/1530 train_loss:3.4920 train_time:168868ms step_avg:170.40ms
step:1002/1530 train_loss:3.3682 train_time:169043ms step_avg:170.41ms
step:1003/1530 train_loss:3.5454 train_time:169220ms step_avg:170.41ms
step:1004/1530 train_loss:3.5941 train_time:169394ms step_avg:170.42ms
step:1005/1530 train_loss:3.3819 train_time:169571ms step_avg:170.42ms
step:1006/1530 train_loss:3.4595 train_time:169748ms step_avg:170.43ms
step:1007/1530 train_loss:3.4260 train_time:169922ms step_avg:170.43ms
step:1008/1530 train_loss:3.5526 train_time:170097ms step_avg:170.44ms
step:1009/1530 train_loss:3.6638 train_time:170276ms step_avg:170.45ms
step:1010/1530 train_loss:3.5537 train_time:170450ms step_avg:170.45ms
step:1011/1530 train_loss:3.5242 train_time:170625ms step_avg:170.45ms
step:1012/1530 train_loss:3.3831 train_time:170799ms step_avg:170.46ms
step:1013/1530 train_loss:3.5277 train_time:170974ms step_avg:170.46ms
step:1014/1530 train_loss:3.6156 train_time:171151ms step_avg:170.47ms
step:1015/1530 train_loss:3.3186 train_time:171328ms step_avg:170.48ms
step:1016/1530 train_loss:3.3997 train_time:171502ms step_avg:170.48ms
step:1017/1530 train_loss:3.3904 train_time:171679ms step_avg:170.49ms
step:1018/1530 train_loss:3.3890 train_time:171854ms step_avg:170.49ms
step:1019/1530 train_loss:3.5141 train_time:172030ms step_avg:170.50ms
step:1020/1530 train_loss:3.3708 train_time:172208ms step_avg:170.50ms
step:1021/1530 train_loss:3.3476 train_time:172383ms step_avg:170.51ms
step:1022/1530 train_loss:3.4667 train_time:172560ms step_avg:170.51ms
step:1023/1530 train_loss:3.4997 train_time:172736ms step_avg:170.52ms
step:1024/1530 train_loss:3.4680 train_time:172912ms step_avg:170.53ms
step:1025/1530 train_loss:3.4714 train_time:173091ms step_avg:170.53ms
step:1026/1530 train_loss:3.6095 train_time:173267ms step_avg:170.54ms
step:1027/1530 train_loss:3.3108 train_time:173443ms step_avg:170.54ms
step:1028/1530 train_loss:3.3879 train_time:173623ms step_avg:170.55ms
step:1029/1530 train_loss:3.3026 train_time:173803ms step_avg:170.56ms
step:1030/1530 train_loss:3.5321 train_time:173980ms step_avg:170.57ms
step:1031/1530 train_loss:3.4989 train_time:174155ms step_avg:170.57ms
step:1032/1530 train_loss:3.6901 train_time:174337ms step_avg:170.58ms
step:1033/1530 train_loss:3.4861 train_time:174513ms step_avg:170.59ms
step:1034/1530 train_loss:3.3905 train_time:174690ms step_avg:170.60ms
step:1035/1530 train_loss:3.4362 train_time:174867ms step_avg:170.60ms
step:1036/1530 train_loss:3.4731 train_time:175044ms step_avg:170.61ms
step:1037/1530 train_loss:3.7821 train_time:175222ms step_avg:170.62ms
step:1038/1530 train_loss:3.6065 train_time:175400ms step_avg:170.62ms
step:1039/1530 train_loss:3.5040 train_time:175582ms step_avg:170.63ms
step:1040/1530 train_loss:3.4030 train_time:175757ms step_avg:170.64ms
step:1041/1530 train_loss:3.4761 train_time:175933ms step_avg:170.64ms
step:1042/1530 train_loss:3.5141 train_time:176107ms step_avg:170.65ms
step:1043/1530 train_loss:3.4397 train_time:176283ms step_avg:170.65ms
step:1044/1530 train_loss:3.4528 train_time:176459ms step_avg:170.66ms
step:1045/1530 train_loss:3.5021 train_time:176638ms step_avg:170.66ms
step:1046/1530 train_loss:3.4184 train_time:176814ms step_avg:170.67ms
step:1047/1530 train_loss:3.6311 train_time:176991ms step_avg:170.68ms
step:1048/1530 train_loss:3.4889 train_time:177169ms step_avg:170.68ms
step:1049/1530 train_loss:3.3924 train_time:177345ms step_avg:170.69ms
step:1050/1530 train_loss:3.3866 train_time:177522ms step_avg:170.69ms
step:1051/1530 train_loss:3.4882 train_time:177699ms step_avg:170.70ms
step:1052/1530 train_loss:3.3524 train_time:177876ms step_avg:170.71ms
step:1053/1530 train_loss:3.6866 train_time:178052ms step_avg:170.71ms
step:1054/1530 train_loss:3.5271 train_time:178229ms step_avg:170.72ms
step:1055/1530 train_loss:3.3802 train_time:178404ms step_avg:170.72ms
step:1056/1530 train_loss:3.4866 train_time:178580ms step_avg:170.73ms
step:1057/1530 train_loss:3.5723 train_time:178756ms step_avg:170.73ms
step:1058/1530 train_loss:3.2963 train_time:178933ms step_avg:170.74ms
step:1059/1530 train_loss:3.3623 train_time:179114ms step_avg:170.75ms
step:1060/1530 train_loss:3.4276 train_time:179291ms step_avg:170.75ms
step:1061/1530 train_loss:3.4107 train_time:179466ms step_avg:170.76ms
step:1062/1530 train_loss:3.3700 train_time:179641ms step_avg:170.76ms
step:1063/1530 train_loss:3.4458 train_time:179815ms step_avg:170.76ms
step:1064/1530 train_loss:3.3748 train_time:179990ms step_avg:170.77ms
step:1065/1530 train_loss:3.3529 train_time:180170ms step_avg:170.78ms
step:1066/1530 train_loss:3.4090 train_time:180347ms step_avg:170.78ms
step:1067/1530 train_loss:3.2684 train_time:180524ms step_avg:170.79ms
step:1068/1530 train_loss:3.4258 train_time:180700ms step_avg:170.79ms
step:1069/1530 train_loss:3.2888 train_time:180879ms step_avg:170.80ms
step:1070/1530 train_loss:3.5616 train_time:181054ms step_avg:170.81ms
step:1071/1530 train_loss:3.5021 train_time:181234ms step_avg:170.81ms
step:1072/1530 train_loss:3.4266 train_time:181410ms step_avg:170.82ms
step:1073/1530 train_loss:3.5185 train_time:181584ms step_avg:170.82ms
step:1074/1530 train_loss:3.4250 train_time:181760ms step_avg:170.83ms
step:1075/1530 train_loss:3.3925 train_time:181936ms step_avg:170.83ms
step:1076/1530 train_loss:3.7946 train_time:182114ms step_avg:170.84ms
step:1077/1530 train_loss:3.4261 train_time:182289ms step_avg:170.84ms
step:1078/1530 train_loss:3.0865 train_time:182475ms step_avg:170.86ms
step:1079/1530 train_loss:3.5276 train_time:182651ms step_avg:170.86ms
step:1080/1530 train_loss:3.4219 train_time:182829ms step_avg:170.87ms
step:1081/1530 train_loss:3.4951 train_time:183003ms step_avg:170.87ms
step:1082/1530 train_loss:3.5837 train_time:183181ms step_avg:170.88ms
step:1083/1530 train_loss:3.4879 train_time:183355ms step_avg:170.88ms
step:1084/1530 train_loss:3.4568 train_time:183531ms step_avg:170.89ms
step:1085/1530 train_loss:3.4252 train_time:183706ms step_avg:170.89ms
step:1086/1530 train_loss:3.6211 train_time:183882ms step_avg:170.89ms
step:1087/1530 train_loss:3.4952 train_time:184057ms step_avg:170.90ms
step:1088/1530 train_loss:3.3649 train_time:184233ms step_avg:170.90ms
step:1089/1530 train_loss:3.3670 train_time:184413ms step_avg:170.91ms
step:1090/1530 train_loss:3.4736 train_time:184592ms step_avg:170.92ms
step:1091/1530 train_loss:3.2772 train_time:184771ms step_avg:170.93ms
step:1092/1530 train_loss:3.4748 train_time:184949ms step_avg:170.93ms
step:1093/1530 train_loss:3.5962 train_time:185126ms step_avg:170.94ms
step:1094/1530 train_loss:3.4390 train_time:185301ms step_avg:170.94ms
step:1095/1530 train_loss:3.4168 train_time:185475ms step_avg:170.94ms
step:1096/1530 train_loss:3.4166 train_time:185651ms step_avg:170.95ms
step:1097/1530 train_loss:3.4839 train_time:185828ms step_avg:170.96ms
step:1098/1530 train_loss:3.5552 train_time:186006ms step_avg:170.96ms
step:1099/1530 train_loss:3.5176 train_time:186185ms step_avg:170.97ms
step:1100/1530 train_loss:3.4203 train_time:186364ms step_avg:170.98ms
step:1101/1530 train_loss:3.2773 train_time:186543ms step_avg:170.98ms
step:1102/1530 train_loss:3.2974 train_time:186721ms step_avg:170.99ms
step:1103/1530 train_loss:3.4379 train_time:186904ms step_avg:171.00ms
step:1104/1530 train_loss:3.3121 train_time:187080ms step_avg:171.01ms
step:1105/1530 train_loss:4.0577 train_time:187258ms step_avg:171.01ms
step:1106/1530 train_loss:3.2154 train_time:187433ms step_avg:171.02ms
step:1107/1530 train_loss:3.5571 train_time:187608ms step_avg:171.02ms
step:1108/1530 train_loss:3.3388 train_time:187783ms step_avg:171.02ms
step:1109/1530 train_loss:3.4950 train_time:187958ms step_avg:171.03ms
step:1110/1530 train_loss:3.4222 train_time:188133ms step_avg:171.03ms
step:1111/1530 train_loss:3.4765 train_time:188309ms step_avg:171.03ms
step:1112/1530 train_loss:3.5510 train_time:188490ms step_avg:171.04ms
step:1113/1530 train_loss:3.4221 train_time:188674ms step_avg:171.06ms
step:1114/1530 train_loss:3.3621 train_time:188855ms step_avg:171.06ms
step:1115/1530 train_loss:3.2330 train_time:189035ms step_avg:171.07ms
step:1116/1530 train_loss:3.4211 train_time:189209ms step_avg:171.08ms
step:1117/1530 train_loss:3.5845 train_time:189389ms step_avg:171.08ms
step:1118/1530 train_loss:3.6146 train_time:189566ms step_avg:171.09ms
step:1119/1530 train_loss:3.4693 train_time:189741ms step_avg:171.09ms
step:1120/1530 train_loss:3.4843 train_time:189916ms step_avg:171.10ms
step:1121/1530 train_loss:3.3837 train_time:190095ms step_avg:171.10ms
step:1122/1530 train_loss:3.4492 train_time:190270ms step_avg:171.11ms
step:1123/1530 train_loss:3.5761 train_time:190445ms step_avg:171.11ms
step:1124/1530 train_loss:3.3277 train_time:190621ms step_avg:171.11ms
step:1125/1530 train_loss:3.2246 train_time:190797ms step_avg:171.12ms
step:1125/1530 val_loss:3.4017 train_time:190847ms step_avg:171.16ms
step:1126/1530 train_loss:3.4660 train_time:190974ms step_avg:171.12ms
step:1127/1530 train_loss:3.6644 train_time:191152ms step_avg:171.13ms
step:1128/1530 train_loss:3.2225 train_time:191330ms step_avg:171.14ms
step:1129/1530 train_loss:3.5505 train_time:191508ms step_avg:171.14ms
step:1130/1530 train_loss:3.3680 train_time:191687ms step_avg:171.15ms
step:1131/1530 train_loss:3.3974 train_time:191870ms step_avg:171.16ms
step:1132/1530 train_loss:3.3606 train_time:192043ms step_avg:171.16ms
step:1133/1530 train_loss:3.4822 train_time:192351ms step_avg:171.28ms
step:1134/1530 train_loss:3.4380 train_time:192536ms step_avg:171.30ms
step:1135/1530 train_loss:3.5165 train_time:192714ms step_avg:171.30ms
step:1136/1530 train_loss:3.5545 train_time:192893ms step_avg:171.31ms
step:1137/1530 train_loss:3.4444 train_time:193069ms step_avg:171.31ms
step:1138/1530 train_loss:3.3460 train_time:193251ms step_avg:171.32ms
step:1139/1530 train_loss:3.6451 train_time:193576ms step_avg:171.46ms
step:1140/1530 train_loss:3.4470 train_time:193753ms step_avg:171.46ms
step:1141/1530 train_loss:3.5898 train_time:193934ms step_avg:171.47ms
step:1142/1530 train_loss:3.4355 train_time:194110ms step_avg:171.48ms
step:1143/1530 train_loss:3.3531 train_time:194288ms step_avg:171.48ms
step:1144/1530 train_loss:3.4385 train_time:194465ms step_avg:171.49ms
step:1145/1530 train_loss:3.5825 train_time:194640ms step_avg:171.49ms
step:1146/1530 train_loss:3.5480 train_time:194821ms step_avg:171.50ms
step:1147/1530 train_loss:3.4745 train_time:194999ms step_avg:171.50ms
step:1148/1530 train_loss:3.4881 train_time:195177ms step_avg:171.51ms
step:1149/1530 train_loss:3.3146 train_time:195357ms step_avg:171.52ms
step:1150/1530 train_loss:3.3652 train_time:195533ms step_avg:171.52ms
step:1151/1530 train_loss:3.3090 train_time:195712ms step_avg:171.53ms
step:1152/1530 train_loss:3.3872 train_time:195894ms step_avg:171.54ms
step:1153/1530 train_loss:3.4217 train_time:196076ms step_avg:171.55ms
step:1154/1530 train_loss:3.5117 train_time:196252ms step_avg:171.55ms
step:1155/1530 train_loss:3.3112 train_time:196437ms step_avg:171.56ms
step:1156/1530 train_loss:3.5277 train_time:196620ms step_avg:171.57ms
step:1157/1530 train_loss:3.4848 train_time:196799ms step_avg:171.58ms
step:1158/1530 train_loss:3.2444 train_time:196976ms step_avg:171.58ms
step:1159/1530 train_loss:3.3428 train_time:197153ms step_avg:171.59ms
step:1160/1530 train_loss:3.3334 train_time:197328ms step_avg:171.59ms
step:1161/1530 train_loss:3.0793 train_time:197506ms step_avg:171.60ms
step:1162/1530 train_loss:3.4129 train_time:197683ms step_avg:171.60ms
step:1163/1530 train_loss:3.3843 train_time:197862ms step_avg:171.61ms
step:1164/1530 train_loss:3.2857 train_time:198038ms step_avg:171.61ms
step:1165/1530 train_loss:3.2402 train_time:198215ms step_avg:171.61ms
step:1166/1530 train_loss:3.3797 train_time:198395ms step_avg:171.62ms
step:1167/1530 train_loss:3.4056 train_time:198571ms step_avg:171.63ms
step:1168/1530 train_loss:3.7158 train_time:198746ms step_avg:171.63ms
step:1169/1530 train_loss:3.3708 train_time:198924ms step_avg:171.63ms
step:1170/1530 train_loss:3.3820 train_time:199101ms step_avg:171.64ms
step:1171/1530 train_loss:3.2971 train_time:199278ms step_avg:171.64ms
step:1172/1530 train_loss:3.4161 train_time:199453ms step_avg:171.65ms
step:1173/1530 train_loss:3.5331 train_time:199634ms step_avg:171.65ms
step:1174/1530 train_loss:3.3750 train_time:199818ms step_avg:171.66ms
step:1175/1530 train_loss:3.3598 train_time:199998ms step_avg:171.67ms
step:1176/1530 train_loss:3.4175 train_time:200181ms step_avg:171.68ms
step:1177/1530 train_loss:3.4420 train_time:200363ms step_avg:171.69ms
step:1178/1530 train_loss:3.4903 train_time:200539ms step_avg:171.69ms
step:1179/1530 train_loss:3.3966 train_time:200715ms step_avg:171.70ms
step:1180/1530 train_loss:3.3466 train_time:200902ms step_avg:171.71ms
step:1181/1530 train_loss:3.3277 train_time:201079ms step_avg:171.72ms
step:1182/1530 train_loss:3.3581 train_time:201257ms step_avg:171.72ms
step:1183/1530 train_loss:3.3240 train_time:201434ms step_avg:171.73ms
step:1184/1530 train_loss:3.5073 train_time:201610ms step_avg:171.73ms
step:1185/1530 train_loss:3.5385 train_time:201792ms step_avg:171.74ms
step:1186/1530 train_loss:3.3586 train_time:201973ms step_avg:171.75ms
step:1187/1530 train_loss:3.4116 train_time:202159ms step_avg:171.76ms
step:1188/1530 train_loss:3.4378 train_time:202335ms step_avg:171.76ms
step:1189/1530 train_loss:3.2730 train_time:202514ms step_avg:171.77ms
step:1190/1530 train_loss:3.4347 train_time:202692ms step_avg:171.77ms
step:1191/1530 train_loss:3.5750 train_time:202871ms step_avg:171.78ms
step:1192/1530 train_loss:3.3852 train_time:203045ms step_avg:171.78ms
step:1193/1530 train_loss:3.2697 train_time:203220ms step_avg:171.78ms
step:1194/1530 train_loss:3.5469 train_time:203397ms step_avg:171.79ms
step:1195/1530 train_loss:3.3638 train_time:203579ms step_avg:171.80ms
step:1196/1530 train_loss:3.3779 train_time:203765ms step_avg:171.81ms
step:1197/1530 train_loss:3.2843 train_time:203945ms step_avg:171.82ms
step:1198/1530 train_loss:3.3009 train_time:204134ms step_avg:171.83ms
step:1199/1530 train_loss:3.3342 train_time:204313ms step_avg:171.84ms
step:1200/1530 train_loss:3.4408 train_time:204491ms step_avg:171.84ms
step:1201/1530 train_loss:3.4752 train_time:204668ms step_avg:171.85ms
step:1202/1530 train_loss:3.6018 train_time:204857ms step_avg:171.86ms
step:1203/1530 train_loss:3.3989 train_time:205036ms step_avg:171.87ms
step:1204/1530 train_loss:3.2997 train_time:205217ms step_avg:171.87ms
step:1205/1530 train_loss:3.4345 train_time:205394ms step_avg:171.88ms
step:1206/1530 train_loss:3.4721 train_time:205570ms step_avg:171.88ms
step:1207/1530 train_loss:3.5118 train_time:205748ms step_avg:171.89ms
step:1208/1530 train_loss:3.3892 train_time:205924ms step_avg:171.89ms
step:1209/1530 train_loss:3.2411 train_time:206104ms step_avg:171.90ms
step:1210/1530 train_loss:3.2969 train_time:206284ms step_avg:171.90ms
step:1211/1530 train_loss:3.3895 train_time:206461ms step_avg:171.91ms
step:1212/1530 train_loss:3.3886 train_time:206640ms step_avg:171.91ms
step:1213/1530 train_loss:3.4014 train_time:206819ms step_avg:171.92ms
step:1214/1530 train_loss:3.2468 train_time:207000ms step_avg:171.93ms
step:1215/1530 train_loss:3.3903 train_time:207177ms step_avg:171.93ms
step:1216/1530 train_loss:3.3214 train_time:207355ms step_avg:171.94ms
step:1217/1530 train_loss:3.3166 train_time:207531ms step_avg:171.94ms
step:1218/1530 train_loss:3.3972 train_time:207708ms step_avg:171.94ms
step:1219/1530 train_loss:3.2480 train_time:207893ms step_avg:171.95ms
step:1220/1530 train_loss:3.4658 train_time:208069ms step_avg:171.96ms
step:1221/1530 train_loss:3.5016 train_time:208243ms step_avg:171.96ms
step:1222/1530 train_loss:3.4267 train_time:208418ms step_avg:171.96ms
step:1223/1530 train_loss:3.2914 train_time:208597ms step_avg:171.97ms
step:1224/1530 train_loss:3.2467 train_time:208780ms step_avg:171.98ms
step:1225/1530 train_loss:3.3587 train_time:208957ms step_avg:171.98ms
step:1226/1530 train_loss:3.3264 train_time:209136ms step_avg:171.99ms
step:1227/1530 train_loss:3.2711 train_time:209316ms step_avg:171.99ms
step:1228/1530 train_loss:3.4425 train_time:209494ms step_avg:172.00ms
step:1229/1530 train_loss:3.3629 train_time:209674ms step_avg:172.00ms
step:1230/1530 train_loss:3.3901 train_time:209856ms step_avg:172.01ms
step:1231/1530 train_loss:3.5685 train_time:210037ms step_avg:172.02ms
step:1232/1530 train_loss:3.4881 train_time:210217ms step_avg:172.03ms
step:1233/1530 train_loss:3.4196 train_time:210394ms step_avg:172.03ms
step:1234/1530 train_loss:3.5789 train_time:210570ms step_avg:172.03ms
step:1235/1530 train_loss:3.3155 train_time:210750ms step_avg:172.04ms
step:1236/1530 train_loss:3.2809 train_time:210928ms step_avg:172.05ms
step:1237/1530 train_loss:3.2657 train_time:211106ms step_avg:172.05ms
step:1238/1530 train_loss:3.2712 train_time:211291ms step_avg:172.06ms
step:1239/1530 train_loss:3.3262 train_time:211467ms step_avg:172.06ms
step:1240/1530 train_loss:3.3776 train_time:211643ms step_avg:172.07ms
step:1241/1530 train_loss:3.4164 train_time:211822ms step_avg:172.07ms
step:1242/1530 train_loss:3.2902 train_time:211999ms step_avg:172.08ms
step:1243/1530 train_loss:3.3958 train_time:212179ms step_avg:172.08ms
step:1244/1530 train_loss:3.4041 train_time:212353ms step_avg:172.08ms
step:1245/1530 train_loss:3.4017 train_time:212530ms step_avg:172.09ms
step:1246/1530 train_loss:3.2361 train_time:212708ms step_avg:172.09ms
step:1247/1530 train_loss:3.3636 train_time:212884ms step_avg:172.10ms
step:1248/1530 train_loss:3.4186 train_time:213061ms step_avg:172.10ms
step:1249/1530 train_loss:3.4147 train_time:213239ms step_avg:172.11ms
step:1250/1530 train_loss:3.2961 train_time:213417ms step_avg:172.11ms
step:1250/1530 val_loss:3.3488 train_time:213471ms step_avg:172.15ms
step:1251/1530 train_loss:3.4839 train_time:213600ms step_avg:172.12ms
step:1252/1530 train_loss:3.3522 train_time:213776ms step_avg:172.12ms
step:1253/1530 train_loss:3.3017 train_time:213952ms step_avg:172.13ms
step:1254/1530 train_loss:3.4124 train_time:214131ms step_avg:172.13ms
step:1255/1530 train_loss:3.5114 train_time:214326ms step_avg:172.15ms
step:1256/1530 train_loss:3.2979 train_time:214509ms step_avg:172.16ms
step:1257/1530 train_loss:3.3719 train_time:214687ms step_avg:172.16ms
step:1258/1530 train_loss:3.3613 train_time:214872ms step_avg:172.17ms
step:1259/1530 train_loss:3.3215 train_time:215051ms step_avg:172.18ms
step:1260/1530 train_loss:3.2075 train_time:215227ms step_avg:172.18ms
step:1261/1530 train_loss:3.2996 train_time:215407ms step_avg:172.19ms
step:1262/1530 train_loss:3.3169 train_time:215589ms step_avg:172.20ms
step:1263/1530 train_loss:3.2330 train_time:215770ms step_avg:172.20ms
step:1264/1530 train_loss:3.4373 train_time:215946ms step_avg:172.21ms
step:1265/1530 train_loss:3.4191 train_time:216122ms step_avg:172.21ms
step:1266/1530 train_loss:3.4362 train_time:216302ms step_avg:172.21ms
step:1267/1530 train_loss:3.3673 train_time:216481ms step_avg:172.22ms
step:1268/1530 train_loss:3.4008 train_time:216660ms step_avg:172.23ms
step:1269/1530 train_loss:3.2495 train_time:216843ms step_avg:172.23ms
step:1270/1530 train_loss:3.1035 train_time:217020ms step_avg:172.24ms
step:1271/1530 train_loss:3.3942 train_time:217197ms step_avg:172.24ms
step:1272/1530 train_loss:3.3483 train_time:217374ms step_avg:172.25ms
step:1273/1530 train_loss:3.3712 train_time:217555ms step_avg:172.25ms
step:1274/1530 train_loss:3.3510 train_time:217736ms step_avg:172.26ms
step:1275/1530 train_loss:3.4295 train_time:217912ms step_avg:172.26ms
step:1276/1530 train_loss:3.4659 train_time:218086ms step_avg:172.26ms
step:1277/1530 train_loss:3.4045 train_time:218265ms step_avg:172.27ms
step:1278/1530 train_loss:3.4029 train_time:218440ms step_avg:172.27ms
step:1279/1530 train_loss:3.2607 train_time:218623ms step_avg:172.28ms
step:1280/1530 train_loss:3.3619 train_time:218809ms step_avg:172.29ms
step:1281/1530 train_loss:3.4190 train_time:218987ms step_avg:172.30ms
step:1282/1530 train_loss:3.4650 train_time:219162ms step_avg:172.30ms
step:1283/1530 train_loss:3.3289 train_time:219340ms step_avg:172.30ms
step:1284/1530 train_loss:3.3667 train_time:219518ms step_avg:172.31ms
step:1285/1530 train_loss:3.3580 train_time:219697ms step_avg:172.31ms
step:1286/1530 train_loss:3.3301 train_time:219875ms step_avg:172.32ms
step:1287/1530 train_loss:3.4838 train_time:220054ms step_avg:172.32ms
step:1288/1530 train_loss:3.2877 train_time:220236ms step_avg:172.33ms
step:1289/1530 train_loss:3.3768 train_time:220422ms step_avg:172.34ms
step:1290/1530 train_loss:3.4490 train_time:220609ms step_avg:172.35ms
step:1291/1530 train_loss:3.3808 train_time:220788ms step_avg:172.36ms
step:1292/1530 train_loss:3.4735 train_time:220970ms step_avg:172.36ms
step:1293/1530 train_loss:3.5100 train_time:221149ms step_avg:172.37ms
step:1294/1530 train_loss:3.4541 train_time:221329ms step_avg:172.37ms
step:1295/1530 train_loss:3.2756 train_time:221508ms step_avg:172.38ms
step:1296/1530 train_loss:3.3688 train_time:221688ms step_avg:172.39ms
step:1297/1530 train_loss:3.2663 train_time:221868ms step_avg:172.39ms
step:1298/1530 train_loss:3.2665 train_time:222050ms step_avg:172.40ms
step:1299/1530 train_loss:3.3934 train_time:222228ms step_avg:172.40ms
step:1300/1530 train_loss:3.3978 train_time:222404ms step_avg:172.41ms
step:1301/1530 train_loss:3.3972 train_time:222580ms step_avg:172.41ms
step:1302/1530 train_loss:3.5696 train_time:222764ms step_avg:172.42ms
step:1303/1530 train_loss:3.2971 train_time:222949ms step_avg:172.43ms
step:1304/1530 train_loss:3.5061 train_time:223130ms step_avg:172.43ms
step:1305/1530 train_loss:3.2535 train_time:223306ms step_avg:172.44ms
step:1306/1530 train_loss:3.4475 train_time:223488ms step_avg:172.44ms
step:1307/1530 train_loss:3.4483 train_time:223664ms step_avg:172.45ms
step:1308/1530 train_loss:3.2829 train_time:223842ms step_avg:172.45ms
step:1309/1530 train_loss:3.3031 train_time:224021ms step_avg:172.46ms
step:1310/1530 train_loss:3.2838 train_time:224197ms step_avg:172.46ms
step:1311/1530 train_loss:3.2919 train_time:224375ms step_avg:172.46ms
step:1312/1530 train_loss:3.3677 train_time:224556ms step_avg:172.47ms
step:1313/1530 train_loss:3.3391 train_time:224732ms step_avg:172.47ms
step:1314/1530 train_loss:3.0397 train_time:224915ms step_avg:172.48ms
step:1315/1530 train_loss:3.2701 train_time:225092ms step_avg:172.48ms
step:1316/1530 train_loss:3.3959 train_time:225267ms step_avg:172.49ms
step:1317/1530 train_loss:3.4176 train_time:225444ms step_avg:172.49ms
step:1318/1530 train_loss:3.2990 train_time:225630ms step_avg:172.50ms
step:1319/1530 train_loss:3.4235 train_time:225810ms step_avg:172.51ms
step:1320/1530 train_loss:3.4529 train_time:225991ms step_avg:172.51ms
step:1321/1530 train_loss:3.3556 train_time:226170ms step_avg:172.52ms
step:1322/1530 train_loss:3.3158 train_time:226478ms step_avg:172.62ms
step:1323/1530 train_loss:3.3190 train_time:226667ms step_avg:172.63ms
step:1324/1530 train_loss:3.4284 train_time:226848ms step_avg:172.64ms
step:1325/1530 train_loss:3.4872 train_time:227032ms step_avg:172.65ms
step:1326/1530 train_loss:3.2062 train_time:227212ms step_avg:172.65ms
step:1327/1530 train_loss:3.1626 train_time:227388ms step_avg:172.66ms
step:1328/1530 train_loss:3.4917 train_time:227567ms step_avg:172.66ms
step:1329/1530 train_loss:3.2943 train_time:227907ms step_avg:172.79ms
step:1330/1530 train_loss:3.4253 train_time:228091ms step_avg:172.80ms
step:1331/1530 train_loss:3.3251 train_time:228267ms step_avg:172.80ms
step:1332/1530 train_loss:3.7401 train_time:228449ms step_avg:172.81ms
step:1333/1530 train_loss:3.4769 train_time:228629ms step_avg:172.81ms
step:1334/1530 train_loss:3.3667 train_time:228808ms step_avg:172.82ms
step:1335/1530 train_loss:3.2875 train_time:228987ms step_avg:172.82ms
step:1336/1530 train_loss:3.2931 train_time:229170ms step_avg:172.83ms
step:1337/1530 train_loss:3.5439 train_time:229349ms step_avg:172.83ms
step:1338/1530 train_loss:3.5166 train_time:229527ms step_avg:172.84ms
step:1339/1530 train_loss:3.3318 train_time:229707ms step_avg:172.84ms
step:1340/1530 train_loss:3.2825 train_time:229886ms step_avg:172.85ms
step:1341/1530 train_loss:3.5912 train_time:230062ms step_avg:172.85ms
step:1342/1530 train_loss:3.3466 train_time:230241ms step_avg:172.85ms
step:1343/1530 train_loss:3.3585 train_time:230420ms step_avg:172.86ms
step:1344/1530 train_loss:3.4094 train_time:230599ms step_avg:172.86ms
step:1345/1530 train_loss:3.3790 train_time:230780ms step_avg:172.87ms
step:1346/1530 train_loss:3.2926 train_time:230957ms step_avg:172.87ms
step:1347/1530 train_loss:3.2759 train_time:231133ms step_avg:172.87ms
step:1348/1530 train_loss:3.3445 train_time:231312ms step_avg:172.88ms
step:1349/1530 train_loss:3.2677 train_time:231489ms step_avg:172.88ms
step:1350/1530 train_loss:3.3824 train_time:231670ms step_avg:172.89ms
step:1351/1530 train_loss:3.2403 train_time:231847ms step_avg:172.89ms
step:1352/1530 train_loss:3.2991 train_time:232025ms step_avg:172.90ms
step:1353/1530 train_loss:3.3994 train_time:232206ms step_avg:172.90ms
step:1354/1530 train_loss:3.2546 train_time:232384ms step_avg:172.90ms
step:1355/1530 train_loss:3.1829 train_time:232559ms step_avg:172.91ms
step:1356/1530 train_loss:3.5074 train_time:232737ms step_avg:172.91ms
step:1357/1530 train_loss:3.4138 train_time:232918ms step_avg:172.92ms
step:1358/1530 train_loss:3.1847 train_time:233097ms step_avg:172.92ms
step:1359/1530 train_loss:3.4361 train_time:233277ms step_avg:172.93ms
step:1360/1530 train_loss:3.3483 train_time:233457ms step_avg:172.93ms
step:1361/1530 train_loss:3.1173 train_time:233643ms step_avg:172.94ms
step:1362/1530 train_loss:3.3881 train_time:233825ms step_avg:172.95ms
step:1363/1530 train_loss:3.2735 train_time:234012ms step_avg:172.96ms
step:1364/1530 train_loss:3.2987 train_time:234190ms step_avg:172.96ms
step:1365/1530 train_loss:3.3085 train_time:234369ms step_avg:172.97ms
step:1366/1530 train_loss:3.4159 train_time:234551ms step_avg:172.97ms
step:1367/1530 train_loss:3.3903 train_time:234730ms step_avg:172.98ms
step:1368/1530 train_loss:3.3389 train_time:234909ms step_avg:172.98ms
step:1369/1530 train_loss:3.2696 train_time:235097ms step_avg:172.99ms
step:1370/1530 train_loss:3.6038 train_time:235277ms step_avg:173.00ms
step:1371/1530 train_loss:3.3106 train_time:235458ms step_avg:173.00ms
step:1372/1530 train_loss:3.3637 train_time:235641ms step_avg:173.01ms
step:1373/1530 train_loss:3.3620 train_time:235818ms step_avg:173.01ms
step:1374/1530 train_loss:3.1442 train_time:236000ms step_avg:173.02ms
step:1375/1530 train_loss:3.5253 train_time:236181ms step_avg:173.03ms
step:1375/1530 val_loss:3.3068 train_time:236231ms step_avg:173.06ms
step:1376/1530 train_loss:3.3406 train_time:236359ms step_avg:173.03ms
step:1377/1530 train_loss:3.4778 train_time:236537ms step_avg:173.03ms
step:1378/1530 train_loss:3.4626 train_time:236715ms step_avg:173.04ms
step:1379/1530 train_loss:3.1152 train_time:236898ms step_avg:173.04ms
step:1380/1530 train_loss:3.3070 train_time:237077ms step_avg:173.05ms
step:1381/1530 train_loss:3.6972 train_time:237261ms step_avg:173.06ms
step:1382/1530 train_loss:3.2021 train_time:237439ms step_avg:173.06ms
step:1383/1530 train_loss:3.3850 train_time:237621ms step_avg:173.07ms
step:1384/1530 train_loss:3.4642 train_time:237803ms step_avg:173.07ms
step:1385/1530 train_loss:3.4012 train_time:237980ms step_avg:173.08ms
step:1386/1530 train_loss:3.3359 train_time:238160ms step_avg:173.08ms
step:1387/1530 train_loss:3.1947 train_time:238339ms step_avg:173.09ms
step:1388/1530 train_loss:3.3448 train_time:238517ms step_avg:173.09ms
step:1389/1530 train_loss:3.3105 train_time:238699ms step_avg:173.10ms
step:1390/1530 train_loss:3.5635 train_time:238876ms step_avg:173.10ms
step:1391/1530 train_loss:3.2877 train_time:239054ms step_avg:173.10ms
step:1392/1530 train_loss:3.2805 train_time:239234ms step_avg:173.11ms
step:1393/1530 train_loss:3.2323 train_time:239413ms step_avg:173.11ms
step:1394/1530 train_loss:3.4935 train_time:239591ms step_avg:173.11ms
step:1395/1530 train_loss:3.3880 train_time:239771ms step_avg:173.12ms
step:1396/1530 train_loss:3.4020 train_time:239948ms step_avg:173.12ms
step:1397/1530 train_loss:3.3073 train_time:240123ms step_avg:173.12ms
step:1398/1530 train_loss:3.2527 train_time:240299ms step_avg:173.13ms
step:1399/1530 train_loss:3.3157 train_time:240479ms step_avg:173.13ms
step:1400/1530 train_loss:3.3120 train_time:240661ms step_avg:173.14ms
step:1401/1530 train_loss:3.3474 train_time:240838ms step_avg:173.14ms
step:1402/1530 train_loss:3.2928 train_time:241018ms step_avg:173.14ms
step:1403/1530 train_loss:3.4874 train_time:241204ms step_avg:173.15ms
step:1404/1530 train_loss:3.2793 train_time:241381ms step_avg:173.16ms
step:1405/1530 train_loss:3.3119 train_time:241562ms step_avg:173.16ms
step:1406/1530 train_loss:3.3086 train_time:241742ms step_avg:173.17ms
step:1407/1530 train_loss:3.1716 train_time:241917ms step_avg:173.17ms
step:1408/1530 train_loss:3.3087 train_time:242097ms step_avg:173.17ms
step:1409/1530 train_loss:3.2949 train_time:242284ms step_avg:173.18ms
step:1410/1530 train_loss:3.2878 train_time:242464ms step_avg:173.19ms
step:1411/1530 train_loss:3.3624 train_time:242638ms step_avg:173.19ms
step:1412/1530 train_loss:3.3292 train_time:242816ms step_avg:173.19ms
step:1413/1530 train_loss:3.3567 train_time:242996ms step_avg:173.20ms
step:1414/1530 train_loss:3.3279 train_time:243177ms step_avg:173.20ms
step:1415/1530 train_loss:3.4078 train_time:243362ms step_avg:173.21ms
step:1416/1530 train_loss:3.2264 train_time:243550ms step_avg:173.22ms
step:1417/1530 train_loss:3.2799 train_time:243734ms step_avg:173.23ms
step:1418/1530 train_loss:3.3843 train_time:243914ms step_avg:173.23ms
step:1419/1530 train_loss:3.3372 train_time:244097ms step_avg:173.24ms
step:1420/1530 train_loss:3.3634 train_time:244279ms step_avg:173.25ms
step:1421/1530 train_loss:3.3663 train_time:244461ms step_avg:173.25ms
step:1422/1530 train_loss:3.3290 train_time:244638ms step_avg:173.26ms
step:1423/1530 train_loss:3.3119 train_time:244817ms step_avg:173.26ms
step:1424/1530 train_loss:3.3298 train_time:245001ms step_avg:173.27ms
step:1425/1530 train_loss:3.1873 train_time:245188ms step_avg:173.28ms
step:1426/1530 train_loss:3.3171 train_time:245367ms step_avg:173.28ms
step:1427/1530 train_loss:3.2814 train_time:245550ms step_avg:173.29ms
step:1428/1530 train_loss:3.3749 train_time:245726ms step_avg:173.29ms
step:1429/1530 train_loss:3.3499 train_time:245904ms step_avg:173.29ms
step:1430/1530 train_loss:3.2537 train_time:246085ms step_avg:173.30ms
step:1431/1530 train_loss:3.3193 train_time:246265ms step_avg:173.30ms
step:1432/1530 train_loss:3.3359 train_time:246449ms step_avg:173.31ms
step:1433/1530 train_loss:3.1288 train_time:246633ms step_avg:173.32ms
step:1434/1530 train_loss:3.2846 train_time:246816ms step_avg:173.33ms
step:1435/1530 train_loss:3.1141 train_time:246996ms step_avg:173.33ms
step:1436/1530 train_loss:3.2246 train_time:247175ms step_avg:173.33ms
step:1437/1530 train_loss:3.4046 train_time:247353ms step_avg:173.34ms
step:1438/1530 train_loss:3.3799 train_time:247530ms step_avg:173.34ms
step:1439/1530 train_loss:3.3128 train_time:247710ms step_avg:173.34ms
step:1440/1530 train_loss:3.1867 train_time:247885ms step_avg:173.35ms
step:1441/1530 train_loss:3.3350 train_time:248065ms step_avg:173.35ms
step:1442/1530 train_loss:3.3791 train_time:248249ms step_avg:173.36ms
step:1443/1530 train_loss:3.4866 train_time:248436ms step_avg:173.37ms
step:1444/1530 train_loss:3.4383 train_time:248613ms step_avg:173.37ms
step:1445/1530 train_loss:3.3366 train_time:248791ms step_avg:173.37ms
step:1446/1530 train_loss:3.1920 train_time:248972ms step_avg:173.38ms
step:1447/1530 train_loss:3.2953 train_time:249153ms step_avg:173.38ms
step:1448/1530 train_loss:3.2942 train_time:249331ms step_avg:173.39ms
step:1449/1530 train_loss:3.3903 train_time:249509ms step_avg:173.39ms
step:1450/1530 train_loss:3.3862 train_time:249691ms step_avg:173.40ms
step:1451/1530 train_loss:3.1972 train_time:249871ms step_avg:173.40ms
step:1452/1530 train_loss:3.3218 train_time:250051ms step_avg:173.41ms
step:1453/1530 train_loss:3.2567 train_time:250227ms step_avg:173.41ms
step:1454/1530 train_loss:3.2832 train_time:250404ms step_avg:173.41ms
step:1455/1530 train_loss:3.3274 train_time:250587ms step_avg:173.42ms
step:1456/1530 train_loss:3.2737 train_time:250765ms step_avg:173.42ms
step:1457/1530 train_loss:3.1522 train_time:250941ms step_avg:173.42ms
step:1458/1530 train_loss:3.4194 train_time:251118ms step_avg:173.42ms
step:1459/1530 train_loss:3.2724 train_time:251301ms step_avg:173.43ms
step:1460/1530 train_loss:3.3048 train_time:251481ms step_avg:173.44ms
step:1461/1530 train_loss:3.4273 train_time:251661ms step_avg:173.44ms
step:1462/1530 train_loss:3.2585 train_time:251837ms step_avg:173.44ms
step:1463/1530 train_loss:3.4641 train_time:252020ms step_avg:173.45ms
step:1464/1530 train_loss:3.3535 train_time:252199ms step_avg:173.45ms
step:1465/1530 train_loss:3.3544 train_time:252379ms step_avg:173.46ms
step:1466/1530 train_loss:3.2874 train_time:252556ms step_avg:173.46ms
step:1467/1530 train_loss:3.3884 train_time:252737ms step_avg:173.46ms
step:1468/1530 train_loss:3.2809 train_time:252914ms step_avg:173.47ms
step:1469/1530 train_loss:3.2710 train_time:253094ms step_avg:173.47ms
step:1470/1530 train_loss:3.3247 train_time:253276ms step_avg:173.48ms
step:1471/1530 train_loss:3.2525 train_time:253462ms step_avg:173.49ms
step:1472/1530 train_loss:3.2425 train_time:253646ms step_avg:173.49ms
step:1473/1530 train_loss:3.4378 train_time:253822ms step_avg:173.49ms
step:1474/1530 train_loss:3.3098 train_time:254007ms step_avg:173.50ms
step:1475/1530 train_loss:3.1458 train_time:254193ms step_avg:173.51ms
step:1476/1530 train_loss:3.2595 train_time:254372ms step_avg:173.51ms
step:1477/1530 train_loss:3.2363 train_time:254558ms step_avg:173.52ms
step:1478/1530 train_loss:3.3003 train_time:254742ms step_avg:173.53ms
step:1479/1530 train_loss:3.3898 train_time:254923ms step_avg:173.54ms
step:1480/1530 train_loss:3.2620 train_time:255103ms step_avg:173.54ms
step:1481/1530 train_loss:3.4501 train_time:255286ms step_avg:173.55ms
step:1482/1530 train_loss:3.3627 train_time:255474ms step_avg:173.56ms
step:1483/1530 train_loss:3.2734 train_time:255662ms step_avg:173.57ms
step:1484/1530 train_loss:3.2619 train_time:255849ms step_avg:173.57ms
step:1485/1530 train_loss:3.2739 train_time:256031ms step_avg:173.58ms
step:1486/1530 train_loss:3.2263 train_time:256214ms step_avg:173.59ms
step:1487/1530 train_loss:3.3366 train_time:256396ms step_avg:173.59ms
step:1488/1530 train_loss:3.2393 train_time:256580ms step_avg:173.60ms
step:1489/1530 train_loss:3.3100 train_time:256761ms step_avg:173.60ms
step:1490/1530 train_loss:3.2489 train_time:256941ms step_avg:173.61ms
step:1491/1530 train_loss:3.1560 train_time:257121ms step_avg:173.61ms
step:1492/1530 train_loss:3.2616 train_time:257302ms step_avg:173.62ms
step:1493/1530 train_loss:3.4317 train_time:257480ms step_avg:173.62ms
step:1494/1530 train_loss:3.2928 train_time:257658ms step_avg:173.62ms
step:1495/1530 train_loss:3.0255 train_time:257841ms step_avg:173.63ms
step:1496/1530 train_loss:3.3572 train_time:258022ms step_avg:173.64ms
step:1497/1530 train_loss:3.3110 train_time:258207ms step_avg:173.64ms
step:1498/1530 train_loss:3.3413 train_time:258390ms step_avg:173.65ms
step:1499/1530 train_loss:3.3113 train_time:258579ms step_avg:173.66ms
step:1500/1530 train_loss:3.2923 train_time:258772ms step_avg:173.67ms
step:1500/1530 val_loss:3.2760 train_time:258826ms step_avg:173.71ms
step:1501/1530 train_loss:3.0832 train_time:258963ms step_avg:173.68ms
step:1502/1530 train_loss:3.3567 train_time:259156ms step_avg:173.70ms
step:1503/1530 train_loss:3.2390 train_time:259334ms step_avg:173.70ms
step:1504/1530 train_loss:3.2487 train_time:259513ms step_avg:173.70ms
step:1505/1530 train_loss:3.2084 train_time:259694ms step_avg:173.71ms
step:1506/1530 train_loss:3.2764 train_time:259875ms step_avg:173.71ms
step:1507/1530 train_loss:3.1761 train_time:260070ms step_avg:173.73ms
step:1508/1530 train_loss:3.4752 train_time:260253ms step_avg:173.73ms
step:1509/1530 train_loss:3.2750 train_time:260431ms step_avg:173.74ms
step:1510/1530 train_loss:3.2653 train_time:260611ms step_avg:173.74ms
step:1511/1530 train_loss:3.4085 train_time:260919ms step_avg:173.83ms
step:1512/1530 train_loss:3.4190 train_time:261106ms step_avg:173.84ms
step:1513/1530 train_loss:3.2668 train_time:261289ms step_avg:173.85ms
step:1514/1530 train_loss:3.0802 train_time:261470ms step_avg:173.85ms
step:1515/1530 train_loss:3.2405 train_time:261652ms step_avg:173.85ms
step:1516/1530 train_loss:3.2520 train_time:261838ms step_avg:173.86ms
step:1517/1530 train_loss:3.2988 train_time:262019ms step_avg:173.87ms
step:1518/1530 train_loss:3.2000 train_time:262202ms step_avg:173.87ms
step:1519/1530 train_loss:3.5034 train_time:262528ms step_avg:173.98ms
step:1520/1530 train_loss:3.1236 train_time:262710ms step_avg:173.98ms
step:1521/1530 train_loss:3.2040 train_time:262887ms step_avg:173.98ms
step:1522/1530 train_loss:3.3538 train_time:263072ms step_avg:173.99ms
step:1523/1530 train_loss:3.2270 train_time:263249ms step_avg:173.99ms
step:1524/1530 train_loss:3.3439 train_time:263429ms step_avg:174.00ms
step:1525/1530 train_loss:3.3339 train_time:263619ms step_avg:174.01ms
step:1526/1530 train_loss:3.2728 train_time:263808ms step_avg:174.02ms
step:1527/1530 train_loss:3.2870 train_time:263988ms step_avg:174.02ms
step:1528/1530 train_loss:3.4063 train_time:264168ms step_avg:174.02ms
step:1529/1530 train_loss:3.4041 train_time:264346ms step_avg:174.03ms
step:1530/1530 train_loss:3.2342 train_time:264523ms step_avg:174.03ms
step:1530/1530 val_loss:3.2736 train_time:264578ms step_avg:174.06ms