shng2025 commited on
Commit
304deea
1 Parent(s): 4d63b0c
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "gpt2",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
 
1
  {
2
+ "_name_or_path": "./",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
gptesla_training.py ADDED
@@ -0,0 +1,286 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+
3
+ import datasets, transformers
4
+ from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer, set_seed
5
+ from transformers.optimization import get_scheduler
6
+ from datasets import load_dataset, DownloadConfig
7
+
8
+ import torch
9
+ from torch.utils.data import IterableDataset
10
+ from torch.utils.data.dataloader import DataLoader
11
+ from torch.utils.tensorboard import SummaryWriter
12
+ from torch.optim import AdamW
13
+
14
+ import logging
15
+ import wandb
16
+ from huggingface_hub import Repository, create_branch
17
+ from accelerate import Accelerator
18
+ from argparse import Namespace
19
+
20
+
21
+ # Set the API token as an environment variable
22
+ os.environ["TOKENIZERS_PARALLELISM"] = "false"
23
+
24
+
25
+ class ConstantLengthDataset(IterableDataset):
26
+
27
+ def __init__(
28
+ self,
29
+ tokenizer,
30
+ dataset,
31
+ seq_length=1024,
32
+ num_of_sequences=1024,
33
+ chars_per_token=3.6,
34
+ ):
35
+ self.tokenizer = tokenizer
36
+ self.concat_token_id = tokenizer.eos_token_id
37
+ self.dataset = dataset
38
+ self.seq_length = seq_length
39
+ self.input_characters = seq_length * chars_per_token * num_of_sequences
40
+
41
+ def __iter__(self):
42
+ iterator = iter(self.dataset)
43
+ more_examples = True
44
+ while more_examples:
45
+ buffer, buffer_len = [], 0
46
+ while True:
47
+ if buffer_len >= self.input_characters:
48
+ m = f"Buffer full: {buffer_len}>={self.input_characters:.0f}"
49
+ # print(m)
50
+ break
51
+ try:
52
+ m = f"Fill buffer: {buffer_len}<{self.input_characters:.0f}"
53
+ # print(m)
54
+ buffer.append(next(iterator)["content"])
55
+ buffer_len += len(buffer[-1])
56
+ except StopIteration:
57
+ # iterator = iter(self.dataset)
58
+ more_examples = False
59
+ break
60
+
61
+ all_token_ids = []
62
+ tokenized_inputs = self.tokenizer(buffer, truncation=False)
63
+ for tokenized_input in tokenized_inputs["input_ids"]:
64
+ all_token_ids.extend(tokenized_input + [self.concat_token_id])
65
+
66
+ for i in range(0, len(all_token_ids), self.seq_length):
67
+ input_ids = all_token_ids[i : i + self.seq_length]
68
+ if len(input_ids) == self.seq_length:
69
+ yield torch.tensor(input_ids)
70
+
71
+
72
+ def setup_logging(project_name):
73
+ logger = logging.getLogger(__name__)
74
+
75
+ dir_name = "./log"
76
+ if not os.path.exists(dir_name):
77
+ os.makedirs(dir_name)
78
+ print(f"Directory '{dir_name}' was created.")
79
+ else:
80
+ print(f"Directory '{dir_name}' already exists.")
81
+
82
+ # setting up log directory
83
+ logging.basicConfig(
84
+ format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
85
+ datefmt="%m/%d/%Y %H:%M:%S",
86
+ level=logging.INFO,
87
+ handlers=[
88
+ logging.FileHandler(f"log/debug_{accelerator.process_index}.log"),
89
+ logging.StreamHandler(),
90
+ ],
91
+ )
92
+ if accelerator.is_main_process: # We only want to set up logging once
93
+ wandb.init(project=project_name, config=args, dir="./../")
94
+ run_name = wandb.run.name
95
+ tb_writer = SummaryWriter()
96
+ tb_writer.add_hparams(vars(args), {"0": 0})
97
+ logger.setLevel(logging.INFO)
98
+ datasets.utils.logging.set_verbosity_debug()
99
+ transformers.utils.logging.set_verbosity_info()
100
+ else:
101
+ tb_writer = None
102
+ run_name = ""
103
+ logger.setLevel(logging.ERROR)
104
+ datasets.utils.logging.set_verbosity_error()
105
+ transformers.utils.logging.set_verbosity_error()
106
+ return logger, tb_writer, run_name
107
+
108
+
109
+ def create_dataloaders(dataset_name):
110
+ train_data = load_dataset(dataset_name + "-train", split="train", streaming=True)
111
+ train_data = train_data.shuffle(buffer_size=args.shuffle_buffer, seed=args.seed)
112
+ valid_data = load_dataset(
113
+ dataset_name + "-valid", split="validation", streaming=True
114
+ )
115
+
116
+ train_dataset = ConstantLengthDataset(
117
+ tokenizer, train_data, seq_length=args.seq_length
118
+ )
119
+ valid_dataset = ConstantLengthDataset(
120
+ tokenizer, valid_data, seq_length=args.seq_length
121
+ )
122
+
123
+ train_dataloader = DataLoader(
124
+ train_dataset, batch_size=args.train_batch_size, num_workers=96
125
+ )
126
+ eval_dataloader = DataLoader(
127
+ valid_dataset, batch_size=args.valid_batch_size, num_workers=1
128
+ )
129
+ return train_dataloader, eval_dataloader
130
+
131
+
132
+ def log_metrics(step, metrics):
133
+ logger.info(f"Step {step}: {metrics}")
134
+ if accelerator.is_main_process:
135
+ wandb.log(metrics)
136
+ [tb_writer.add_scalar(k, v, step) for k, v in metrics.items()]
137
+
138
+
139
+ def get_grouped_params(model, no_decay=["bias", "LayerNorm.weight"]):
140
+ params_with_wd, params_without_wd = [], []
141
+ for n, p in model.named_parameters():
142
+ if any(nd in n for nd in no_decay):
143
+ params_without_wd.append(p)
144
+ else:
145
+ params_with_wd.append(p)
146
+ return [
147
+ {"params": params_with_wd, "weight_decay": args.weight_decay},
148
+ {"params": params_without_wd, "weight_decay": 0.0},
149
+ ]
150
+
151
+
152
+ def evaluate():
153
+ model.eval()
154
+ losses = []
155
+ for step, batch in enumerate(eval_dataloader):
156
+ with torch.no_grad():
157
+ outputs = model(batch, labels=batch)
158
+ loss = outputs.loss.repeat(args.valid_batch_size)
159
+ losses.append(accelerator.gather(loss))
160
+ if args.max_eval_steps > 0 and step >= args.max_eval_steps:
161
+ break
162
+ loss = torch.mean(torch.cat(losses))
163
+
164
+ try:
165
+ perplexity = torch.exp(loss)
166
+ except OverflowError:
167
+ perplexity = torch.tensor(float("inf"))
168
+
169
+ return loss.item(), perplexity.item()
170
+
171
+
172
+ # Accelerator
173
+ accelerator = Accelerator(dispatch_batches=True)
174
+ acc_state = {str(k): str(v) for k, v in accelerator.state.__dict__.items()}
175
+
176
+ project_name = "shng2025/gptesla-small"
177
+ dataset_name = "shng2025/gptesla"
178
+
179
+ # GPTesla - 111M param setup in comment. Modification to make lighter training requirement needed
180
+ config = {
181
+ "train_batch_size": 12, # 12
182
+ "valid_batch_size": 12, # 12
183
+ "weight_decay": 0.1,
184
+ "shuffle_buffer": 1000,
185
+ "learning_rate": 5e-4, # 5e-4
186
+ "lr_scheduler_type": "cosine",
187
+ "num_warmup_steps": 700, # 2000
188
+ "gradient_accumulation_steps": 1, # 1
189
+ "max_train_steps": 50000, # 150000
190
+ "max_eval_steps": 10,
191
+ "seq_length": 1024,
192
+ "seed": 1,
193
+ "save_checkpoint_steps": 50,
194
+ } # 15000
195
+
196
+ args = Namespace(**config, **acc_state)
197
+ samples_per_step = accelerator.state.num_processes * args.train_batch_size
198
+ set_seed(args.seed)
199
+
200
+ # Logging
201
+ logger, tb_writer, run_name = setup_logging(project_name.split("/")[1])
202
+ logger.info(accelerator.state)
203
+
204
+ # Load model and tokenizer
205
+ if accelerator.is_main_process:
206
+ new_branch_name = run_name
207
+ create_branch(project_name, repo_type="model", branch=new_branch_name)
208
+ hf_repo = Repository("./", clone_from=project_name, revision=run_name)
209
+
210
+ model = AutoModelForCausalLM.from_pretrained("./") # , gradient_checkpointing=True)
211
+ tokenizer = AutoTokenizer.from_pretrained("./")
212
+
213
+ # Load dataset and dataloader
214
+ train_dataloader, eval_dataloader = create_dataloaders(dataset_name)
215
+
216
+ # Prepare the optimizer and learning rate scheduler
217
+ optimizer = AdamW(get_grouped_params(model), lr=args.learning_rate)
218
+ lr_scheduler = get_scheduler(
219
+ name=args.lr_scheduler_type,
220
+ optimizer=optimizer,
221
+ num_warmup_steps=args.num_warmup_steps,
222
+ num_training_steps=args.max_train_steps,
223
+ )
224
+
225
+
226
+ def get_lr():
227
+ return optimizer.param_groups[0]["lr"]
228
+
229
+
230
+ # Prepare everything with our `accelerator` (order of args is not important)
231
+ model, optimizer, train_dataloader, eval_dataloader = accelerator.prepare(
232
+ model, optimizer, train_dataloader, eval_dataloader
233
+ )
234
+
235
+ # Train model
236
+ model.train()
237
+ completed_steps = 0
238
+ for step, batch in enumerate(train_dataloader, start=1):
239
+ loss = model(batch, labels=batch).loss
240
+ log_metrics(
241
+ step,
242
+ {
243
+ "lr": get_lr(),
244
+ "samples": step * samples_per_step,
245
+ "steps": completed_steps,
246
+ "loss/train": loss.item(),
247
+ },
248
+ )
249
+ loss = loss / args.gradient_accumulation_steps
250
+ accelerator.backward(loss)
251
+ if step % args.gradient_accumulation_steps == 0:
252
+ optimizer.step()
253
+ lr_scheduler.step()
254
+ optimizer.zero_grad()
255
+ completed_steps += 1
256
+ if step % args.save_checkpoint_steps == 0:
257
+ logger.info("Evaluating and saving model checkpoint")
258
+ eval_loss, perplexity = evaluate()
259
+ log_metrics(step, {"loss/eval": eval_loss, "perplexity": perplexity})
260
+ accelerator.wait_for_everyone()
261
+ unwrapped_model = accelerator.unwrap_model(model)
262
+ if accelerator.is_main_process:
263
+ # print(model.state_dict())
264
+ # print(optimizer.state_dict())
265
+ # print(steps)
266
+ # print(train_dataloader.state_dict())
267
+ # print(eval_dataloader.state_dict())
268
+ worker_info = torch.utils.data.get_worker_info()
269
+ print(worker_info)
270
+ unwrapped_model.save_pretrained("./")
271
+ accelerator.save_state(output_dir="my_checkpoint")
272
+ hf_repo.push_to_hub(commit_message=f"step {step}")
273
+ model.train()
274
+ if completed_steps >= args.max_train_steps:
275
+ break
276
+
277
+
278
+ # Evaluate and save the last checkpoint
279
+ logger.info("Evaluating and saving model after training")
280
+ eval_loss, perplexity = evaluate()
281
+ log_metrics(step, {"loss/eval": eval_loss, "perplexity": perplexity})
282
+ accelerator.wait_for_everyone()
283
+ unwrapped_model = accelerator.unwrap_model(model)
284
+ if accelerator.is_main_process:
285
+ unwrapped_model.save_pretrained("./")
286
+ hf_repo.push_to_hub(commit_message="final model")
log/debug_0.log CHANGED
@@ -0,0 +1,409 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 07/25/2024 06:16:39 - INFO - __main__ - Distributed environment: MULTI_GPU Backend: nccl
2
+ Num processes: 4
3
+ Process index: 0
4
+ Local process index: 0
5
+ Device: cuda:0
6
+
7
+ Mixed precision type: fp16
8
+
9
+ 07/25/2024 06:16:39 - WARNING - huggingface_hub.repository - /dli/gptesla-small/./ is already a clone of https://huggingface.co/shng2025/gptesla-small. Make sure you pull the latest changes with `repo.git_pull()`.
10
+ 07/25/2024 06:16:40 - WARNING - huggingface_hub.repository - Revision `hopeful-snow-127` does not exist. Created and checked out branch `hopeful-snow-127`.
11
+ 07/25/2024 06:16:40 - WARNING - huggingface_hub.repository -
12
+ 07/25/2024 06:16:41 - DEBUG - datasets.utils._dataset_viewer - Dataset info for shng2025/gptesla-train is not completely ready yet.
13
+ 07/25/2024 06:16:41 - INFO - datasets.builder - No config specified, defaulting to the single config: gptesla-train/default
14
+ 07/25/2024 06:16:41 - INFO - datasets.info - Loading Dataset Infos from /usr/local/lib/python3.10/dist-packages/datasets/packaged_modules/json
15
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#0, ': Starting to iterate over 2/183 shards.
16
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#1, ': Starting to iterate over 2/183 shards.
17
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#2, ': Starting to iterate over 2/183 shards.
18
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#3, ': Starting to iterate over 2/183 shards.
19
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#4, ': Starting to iterate over 2/183 shards.
20
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#5, ': Starting to iterate over 2/183 shards.
21
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#7, ': Starting to iterate over 2/183 shards.
22
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#8, ': Starting to iterate over 2/183 shards.
23
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#9, ': Starting to iterate over 2/183 shards.
24
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#6, ': Starting to iterate over 2/183 shards.
25
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#10, ': Starting to iterate over 2/183 shards.
26
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#11, ': Starting to iterate over 2/183 shards.
27
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#12, ': Starting to iterate over 2/183 shards.
28
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#13, ': Starting to iterate over 2/183 shards.
29
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#14, ': Starting to iterate over 2/183 shards.
30
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#15, ': Starting to iterate over 2/183 shards.
31
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#16, ': Starting to iterate over 2/183 shards.
32
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#17, ': Starting to iterate over 2/183 shards.
33
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#18, ': Starting to iterate over 2/183 shards.
34
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#19, ': Starting to iterate over 2/183 shards.
35
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#20, ': Starting to iterate over 2/183 shards.
36
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#21, ': Starting to iterate over 2/183 shards.
37
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#22, ': Starting to iterate over 2/183 shards.
38
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#24, ': Starting to iterate over 2/183 shards.
39
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#23, ': Starting to iterate over 2/183 shards.
40
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#25, ': Starting to iterate over 2/183 shards.
41
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#26, ': Starting to iterate over 2/183 shards.
42
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#27, ': Starting to iterate over 2/183 shards.
43
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#28, ': Starting to iterate over 2/183 shards.
44
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#29, ': Starting to iterate over 2/183 shards.
45
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#31, ': Starting to iterate over 2/183 shards.
46
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#30, ': Starting to iterate over 2/183 shards.
47
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#32, ': Starting to iterate over 2/183 shards.
48
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#33, ': Starting to iterate over 2/183 shards.
49
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#34, ': Starting to iterate over 2/183 shards.
50
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#35, ': Starting to iterate over 2/183 shards.
51
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#36, ': Starting to iterate over 2/183 shards.
52
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#37, ': Starting to iterate over 2/183 shards.
53
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#38, ': Starting to iterate over 2/183 shards.
54
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#39, ': Starting to iterate over 2/183 shards.
55
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#40, ': Starting to iterate over 2/183 shards.
56
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#41, ': Starting to iterate over 2/183 shards.
57
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#43, ': Starting to iterate over 2/183 shards.
58
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#44, ': Starting to iterate over 2/183 shards.
59
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#45, ': Starting to iterate over 2/183 shards.
60
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#46, ': Starting to iterate over 2/183 shards.
61
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#42, ': Starting to iterate over 2/183 shards.
62
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#47, ': Starting to iterate over 2/183 shards.
63
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#48, ': Starting to iterate over 2/183 shards.
64
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#49, ': Starting to iterate over 2/183 shards.
65
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#50, ': Starting to iterate over 2/183 shards.
66
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#51, ': Starting to iterate over 2/183 shards.
67
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#52, ': Starting to iterate over 2/183 shards.
68
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#54, ': Starting to iterate over 2/183 shards.
69
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#53, ': Starting to iterate over 2/183 shards.
70
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#55, ': Starting to iterate over 2/183 shards.
71
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#56, ': Starting to iterate over 2/183 shards.
72
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#57, ': Starting to iterate over 2/183 shards.
73
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#58, ': Starting to iterate over 2/183 shards.
74
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#60, ': Starting to iterate over 2/183 shards.
75
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#59, ': Starting to iterate over 2/183 shards.
76
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#61, ': Starting to iterate over 2/183 shards.
77
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#62, ': Starting to iterate over 2/183 shards.
78
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#63, ': Starting to iterate over 2/183 shards.
79
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#64, ': Starting to iterate over 2/183 shards.
80
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#65, ': Starting to iterate over 2/183 shards.
81
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#66, ': Starting to iterate over 2/183 shards.
82
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#68, ': Starting to iterate over 2/183 shards.
83
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#67, ': Starting to iterate over 2/183 shards.
84
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#69, ': Starting to iterate over 2/183 shards.
85
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#70, ': Starting to iterate over 2/183 shards.
86
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#71, ': Starting to iterate over 2/183 shards.
87
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#72, ': Starting to iterate over 2/183 shards.
88
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#73, ': Starting to iterate over 2/183 shards.
89
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#74, ': Starting to iterate over 2/183 shards.
90
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#75, ': Starting to iterate over 2/183 shards.
91
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#77, ': Starting to iterate over 2/183 shards.
92
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#78, ': Starting to iterate over 2/183 shards.
93
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#76, ': Starting to iterate over 2/183 shards.
94
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#79, ': Starting to iterate over 2/183 shards.
95
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#80, ': Starting to iterate over 2/183 shards.
96
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#81, ': Starting to iterate over 2/183 shards.
97
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#82, ': Starting to iterate over 2/183 shards.
98
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#83, ': Starting to iterate over 2/183 shards.
99
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#84, ': Starting to iterate over 2/183 shards.
100
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#85, ': Starting to iterate over 2/183 shards.
101
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#86, ': Starting to iterate over 2/183 shards.
102
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#87, ': Starting to iterate over 1/183 shards.
103
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#88, ': Starting to iterate over 1/183 shards.
104
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#89, ': Starting to iterate over 1/183 shards.
105
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#90, ': Starting to iterate over 1/183 shards.
106
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#91, ': Starting to iterate over 1/183 shards.
107
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#92, ': Starting to iterate over 1/183 shards.
108
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#94, ': Starting to iterate over 1/183 shards.
109
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#93, ': Starting to iterate over 1/183 shards.
110
+ 07/25/2024 06:16:47 - DEBUG - datasets.iterable_dataset - dataloader worker#95, ': Starting to iterate over 1/183 shards.
111
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491327 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
112
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489635 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
113
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497218 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
114
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10500930 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
115
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10621496 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
116
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10668116 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
117
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489599 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
118
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492277 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
119
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495973 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
120
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
121
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
122
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511604 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
123
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
124
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491889 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
125
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488608 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
126
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
127
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10562022 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
128
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
129
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
130
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
131
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487790 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
132
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10863935 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
133
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
134
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
135
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
136
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525688 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
137
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488098 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
138
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
139
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525926 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
140
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491272 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
141
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497335 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
142
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
143
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509262 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
144
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486397 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
145
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10493913 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
146
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10515063 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
147
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10751338 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
148
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488150 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
149
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10949076 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
150
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
151
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
152
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10501535 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
153
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
154
+ 07/25/2024 06:16:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
155
+ 07/25/2024 06:16:49 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
156
+ 07/25/2024 06:16:49 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
157
+ 07/25/2024 06:16:54 - INFO - __main__ - Step 1: {'lr': 0.0, 'samples': 48, 'steps': 0, 'loss/train': 10.554669380187988}
158
+ 07/25/2024 06:16:55 - INFO - __main__ - Step 2: {'lr': 7.142857142857143e-07, 'samples': 96, 'steps': 1, 'loss/train': 10.494059562683105}
159
+ 07/25/2024 06:22:39 - INFO - __main__ - Distributed environment: MULTI_GPU Backend: nccl
160
+ Num processes: 4
161
+ Process index: 0
162
+ Local process index: 0
163
+ Device: cuda:0
164
+
165
+ Mixed precision type: fp16
166
+
167
+ 07/25/2024 06:22:39 - WARNING - huggingface_hub.repository - /dli/gptesla-small/./ is already a clone of https://huggingface.co/shng2025/gptesla-small. Make sure you pull the latest changes with `repo.git_pull()`.
168
+ 07/25/2024 06:22:39 - WARNING - huggingface_hub.repository - Revision `celestial-aardvark-128` does not exist. Created and checked out branch `celestial-aardvark-128`.
169
+ 07/25/2024 06:22:39 - WARNING - huggingface_hub.repository -
170
+ 07/25/2024 06:22:41 - DEBUG - datasets.utils._dataset_viewer - Dataset info for shng2025/gptesla-train is not completely ready yet.
171
+ 07/25/2024 06:22:41 - INFO - datasets.builder - No config specified, defaulting to the single config: gptesla-train/default
172
+ 07/25/2024 06:22:41 - INFO - datasets.info - Loading Dataset Infos from /usr/local/lib/python3.10/dist-packages/datasets/packaged_modules/json
173
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#0, ': Starting to iterate over 2/183 shards.
174
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#1, ': Starting to iterate over 2/183 shards.
175
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#4, ': Starting to iterate over 2/183 shards.
176
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#5, ': Starting to iterate over 2/183 shards.
177
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#2, ': Starting to iterate over 2/183 shards.
178
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#3, ': Starting to iterate over 2/183 shards.
179
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#6, ': Starting to iterate over 2/183 shards.
180
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#7, ': Starting to iterate over 2/183 shards.
181
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#8, ': Starting to iterate over 2/183 shards.
182
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#9, ': Starting to iterate over 2/183 shards.
183
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#10, ': Starting to iterate over 2/183 shards.
184
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#12, ': Starting to iterate over 2/183 shards.
185
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#15, ': Starting to iterate over 2/183 shards.
186
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#14, ': Starting to iterate over 2/183 shards.
187
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#16, ': Starting to iterate over 2/183 shards.
188
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#13, ': Starting to iterate over 2/183 shards.
189
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#11, ': Starting to iterate over 2/183 shards.
190
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#17, ': Starting to iterate over 2/183 shards.
191
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#18, ': Starting to iterate over 2/183 shards.
192
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#19, ': Starting to iterate over 2/183 shards.
193
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#20, ': Starting to iterate over 2/183 shards.
194
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#22, ': Starting to iterate over 2/183 shards.
195
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#23, ': Starting to iterate over 2/183 shards.
196
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#24, ': Starting to iterate over 2/183 shards.
197
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#25, ': Starting to iterate over 2/183 shards.
198
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#21, ': Starting to iterate over 2/183 shards.
199
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#28, ': Starting to iterate over 2/183 shards.
200
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#27, ': Starting to iterate over 2/183 shards.
201
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#26, ': Starting to iterate over 2/183 shards.
202
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#29, ': Starting to iterate over 2/183 shards.
203
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#30, ': Starting to iterate over 2/183 shards.
204
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#31, ': Starting to iterate over 2/183 shards.
205
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#32, ': Starting to iterate over 2/183 shards.
206
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#33, ': Starting to iterate over 2/183 shards.
207
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#34, ': Starting to iterate over 2/183 shards.
208
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#35, ': Starting to iterate over 2/183 shards.
209
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#36, ': Starting to iterate over 2/183 shards.
210
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#37, ': Starting to iterate over 2/183 shards.
211
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#38, ': Starting to iterate over 2/183 shards.
212
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#39, ': Starting to iterate over 2/183 shards.
213
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#40, ': Starting to iterate over 2/183 shards.
214
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#41, ': Starting to iterate over 2/183 shards.
215
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#42, ': Starting to iterate over 2/183 shards.
216
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#43, ': Starting to iterate over 2/183 shards.
217
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#44, ': Starting to iterate over 2/183 shards.
218
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#45, ': Starting to iterate over 2/183 shards.
219
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#46, ': Starting to iterate over 2/183 shards.
220
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#47, ': Starting to iterate over 2/183 shards.
221
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#48, ': Starting to iterate over 2/183 shards.
222
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#49, ': Starting to iterate over 2/183 shards.
223
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#50, ': Starting to iterate over 2/183 shards.
224
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#52, ': Starting to iterate over 2/183 shards.
225
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#53, ': Starting to iterate over 2/183 shards.
226
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#54, ': Starting to iterate over 2/183 shards.
227
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#51, ': Starting to iterate over 2/183 shards.
228
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#55, ': Starting to iterate over 2/183 shards.
229
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#56, ': Starting to iterate over 2/183 shards.
230
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#57, ': Starting to iterate over 2/183 shards.
231
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#58, ': Starting to iterate over 2/183 shards.
232
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#59, ': Starting to iterate over 2/183 shards.
233
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#60, ': Starting to iterate over 2/183 shards.
234
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#61, ': Starting to iterate over 2/183 shards.
235
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#62, ': Starting to iterate over 2/183 shards.
236
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#63, ': Starting to iterate over 2/183 shards.
237
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#64, ': Starting to iterate over 2/183 shards.
238
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#65, ': Starting to iterate over 2/183 shards.
239
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#67, ': Starting to iterate over 2/183 shards.
240
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#66, ': Starting to iterate over 2/183 shards.
241
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#68, ': Starting to iterate over 2/183 shards.
242
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#69, ': Starting to iterate over 2/183 shards.
243
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#70, ': Starting to iterate over 2/183 shards.
244
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#72, ': Starting to iterate over 2/183 shards.
245
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#73, ': Starting to iterate over 2/183 shards.
246
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#74, ': Starting to iterate over 2/183 shards.
247
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#75, ': Starting to iterate over 2/183 shards.
248
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#76, ': Starting to iterate over 2/183 shards.
249
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#71, ': Starting to iterate over 2/183 shards.
250
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#77, ': Starting to iterate over 2/183 shards.
251
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#78, ': Starting to iterate over 2/183 shards.
252
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#79, ': Starting to iterate over 2/183 shards.
253
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#80, ': Starting to iterate over 2/183 shards.
254
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#81, ': Starting to iterate over 2/183 shards.
255
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#82, ': Starting to iterate over 2/183 shards.
256
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#83, ': Starting to iterate over 2/183 shards.
257
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#84, ': Starting to iterate over 2/183 shards.
258
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#85, ': Starting to iterate over 2/183 shards.
259
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#86, ': Starting to iterate over 2/183 shards.
260
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#87, ': Starting to iterate over 1/183 shards.
261
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#88, ': Starting to iterate over 1/183 shards.
262
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#89, ': Starting to iterate over 1/183 shards.
263
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#90, ': Starting to iterate over 1/183 shards.
264
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#92, ': Starting to iterate over 1/183 shards.
265
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#91, ': Starting to iterate over 1/183 shards.
266
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#93, ': Starting to iterate over 1/183 shards.
267
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#94, ': Starting to iterate over 1/183 shards.
268
+ 07/25/2024 06:22:46 - DEBUG - datasets.iterable_dataset - dataloader worker#95, ': Starting to iterate over 1/183 shards.
269
+ 07/25/2024 06:22:46 - DEBUG - datasets.packaged_modules.json.json - Batch of 10500930 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
270
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
271
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492277 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
272
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525688 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
273
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489635 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
274
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486023 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
275
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
276
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
277
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10668116 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
278
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10522596 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
279
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10512203 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
280
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492861 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
281
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497218 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
282
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485912 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
283
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486397 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
284
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10536479 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
285
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10863935 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
286
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491327 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
287
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10562022 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
288
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
289
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485842 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
290
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497111 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
291
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489599 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
292
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
293
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10493913 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
294
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10949076 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
295
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10553677 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
296
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10598254 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
297
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10553677 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
298
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10515063 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
299
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509286 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
300
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487790 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
301
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485847 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
302
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488385 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
303
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10610581 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
304
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495973 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
305
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497062 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
306
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488098 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
307
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511500 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
308
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
309
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10525926 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
310
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488150 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
311
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
312
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486801 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
313
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488651 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
314
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
315
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10499106 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
316
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10552417 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
317
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486616 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
318
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491272 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
319
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511604 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
320
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 11286262 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
321
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491889 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
322
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487725 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
323
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486276 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
324
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 11286262 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
325
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10488608 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
326
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10501535 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
327
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10497335 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
328
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10509262 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
329
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10489575 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
330
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10485918 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
331
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10491547 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
332
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
333
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487097 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
334
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10495520 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
335
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10751338 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
336
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10621496 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
337
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10498167 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
338
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10486172 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
339
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10686322 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
340
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10499607 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
341
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10511515 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
342
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 11115863 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
343
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10530453 bytes couldn't be parsed with block_size=655360. Retrying with block_size=1310720.
344
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10492554 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
345
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10640425 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
346
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10487482 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
347
+ 07/25/2024 06:22:47 - DEBUG - datasets.packaged_modules.json.json - Batch of 10500290 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
348
+ 07/25/2024 06:22:48 - DEBUG - datasets.packaged_modules.json.json - Batch of 10676628 bytes couldn't be parsed with block_size=327680. Retrying with block_size=655360.
349
+ 07/25/2024 06:22:59 - INFO - __main__ - Step 1: {'lr': 0.0, 'samples': 48, 'steps': 0, 'loss/train': 10.554669380187988}
350
+ 07/25/2024 06:23:02 - INFO - __main__ - Step 2: {'lr': 7.142857142857143e-07, 'samples': 96, 'steps': 1, 'loss/train': 10.494059562683105}
351
+ 07/25/2024 06:23:02 - INFO - __main__ - Step 3: {'lr': 1.4285714285714286e-06, 'samples': 144, 'steps': 2, 'loss/train': 10.507988929748535}
352
+ 07/25/2024 06:23:03 - INFO - __main__ - Step 4: {'lr': 2.142857142857143e-06, 'samples': 192, 'steps': 3, 'loss/train': 10.415447235107422}
353
+ 07/25/2024 06:23:03 - INFO - __main__ - Step 5: {'lr': 2.8571428571428573e-06, 'samples': 240, 'steps': 4, 'loss/train': 10.345850944519043}
354
+ 07/25/2024 06:23:03 - INFO - __main__ - Step 6: {'lr': 3.5714285714285714e-06, 'samples': 288, 'steps': 5, 'loss/train': 10.195524215698242}
355
+ 07/25/2024 06:23:03 - INFO - __main__ - Step 7: {'lr': 4.285714285714286e-06, 'samples': 336, 'steps': 6, 'loss/train': 10.09341812133789}
356
+ 07/25/2024 06:23:04 - INFO - __main__ - Step 8: {'lr': 5e-06, 'samples': 384, 'steps': 7, 'loss/train': 9.965239524841309}
357
+ 07/25/2024 06:23:04 - INFO - __main__ - Step 9: {'lr': 5.7142857142857145e-06, 'samples': 432, 'steps': 8, 'loss/train': 9.698853492736816}
358
+ 07/25/2024 06:23:04 - INFO - __main__ - Step 10: {'lr': 6.428571428571429e-06, 'samples': 480, 'steps': 9, 'loss/train': 9.80683708190918}
359
+ 07/25/2024 06:23:05 - INFO - __main__ - Step 11: {'lr': 7.142857142857143e-06, 'samples': 528, 'steps': 10, 'loss/train': 9.633079528808594}
360
+ 07/25/2024 06:23:05 - INFO - __main__ - Step 12: {'lr': 7.857142857142858e-06, 'samples': 576, 'steps': 11, 'loss/train': 9.700591087341309}
361
+ 07/25/2024 06:23:05 - INFO - __main__ - Step 13: {'lr': 8.571428571428573e-06, 'samples': 624, 'steps': 12, 'loss/train': 9.603139877319336}
362
+ 07/25/2024 06:23:05 - INFO - __main__ - Step 14: {'lr': 9.285714285714286e-06, 'samples': 672, 'steps': 13, 'loss/train': 9.30308723449707}
363
+ 07/25/2024 06:23:06 - INFO - __main__ - Step 15: {'lr': 1e-05, 'samples': 720, 'steps': 14, 'loss/train': 9.333526611328125}
364
+ 07/25/2024 06:23:06 - INFO - __main__ - Step 16: {'lr': 1.0714285714285714e-05, 'samples': 768, 'steps': 15, 'loss/train': 8.336181640625}
365
+ 07/25/2024 06:23:06 - INFO - __main__ - Step 17: {'lr': 1.1428571428571429e-05, 'samples': 816, 'steps': 16, 'loss/train': 9.075631141662598}
366
+ 07/25/2024 06:23:07 - INFO - __main__ - Step 18: {'lr': 1.2142857142857142e-05, 'samples': 864, 'steps': 17, 'loss/train': 9.18478012084961}
367
+ 07/25/2024 06:23:07 - INFO - __main__ - Step 19: {'lr': 1.2857142857142857e-05, 'samples': 912, 'steps': 18, 'loss/train': 8.96328353881836}
368
+ 07/25/2024 06:23:07 - INFO - __main__ - Step 20: {'lr': 1.3571428571428572e-05, 'samples': 960, 'steps': 19, 'loss/train': 9.45018196105957}
369
+ 07/25/2024 06:23:07 - INFO - __main__ - Step 21: {'lr': 1.4285714285714285e-05, 'samples': 1008, 'steps': 20, 'loss/train': 8.517333984375}
370
+ 07/25/2024 06:23:08 - INFO - __main__ - Step 22: {'lr': 1.5e-05, 'samples': 1056, 'steps': 21, 'loss/train': 9.207684516906738}
371
+ 07/25/2024 06:23:08 - INFO - __main__ - Step 23: {'lr': 1.5714285714285715e-05, 'samples': 1104, 'steps': 22, 'loss/train': 8.681092262268066}
372
+ 07/25/2024 06:23:08 - INFO - __main__ - Step 24: {'lr': 1.642857142857143e-05, 'samples': 1152, 'steps': 23, 'loss/train': 8.316036224365234}
373
+ 07/25/2024 06:23:09 - INFO - __main__ - Step 25: {'lr': 1.7142857142857145e-05, 'samples': 1200, 'steps': 24, 'loss/train': 8.944169044494629}
374
+ 07/25/2024 06:23:09 - INFO - __main__ - Step 26: {'lr': 1.7857142857142855e-05, 'samples': 1248, 'steps': 25, 'loss/train': 8.878201484680176}
375
+ 07/25/2024 06:23:09 - INFO - __main__ - Step 27: {'lr': 1.8571428571428572e-05, 'samples': 1296, 'steps': 26, 'loss/train': 9.158102989196777}
376
+ 07/25/2024 06:23:09 - INFO - __main__ - Step 28: {'lr': 1.9285714285714285e-05, 'samples': 1344, 'steps': 27, 'loss/train': 9.14354419708252}
377
+ 07/25/2024 06:23:10 - INFO - __main__ - Step 29: {'lr': 2e-05, 'samples': 1392, 'steps': 28, 'loss/train': 8.860624313354492}
378
+ 07/25/2024 06:23:10 - INFO - __main__ - Step 30: {'lr': 2.0714285714285715e-05, 'samples': 1440, 'steps': 29, 'loss/train': 8.876450538635254}
379
+ 07/25/2024 06:23:10 - INFO - __main__ - Step 31: {'lr': 2.1428571428571428e-05, 'samples': 1488, 'steps': 30, 'loss/train': 8.425738334655762}
380
+ 07/25/2024 06:23:10 - INFO - __main__ - Step 32: {'lr': 2.214285714285714e-05, 'samples': 1536, 'steps': 31, 'loss/train': 8.942279815673828}
381
+ 07/25/2024 06:23:11 - INFO - __main__ - Step 33: {'lr': 2.2857142857142858e-05, 'samples': 1584, 'steps': 32, 'loss/train': 8.757084846496582}
382
+ 07/25/2024 06:23:11 - INFO - __main__ - Step 34: {'lr': 2.3571428571428575e-05, 'samples': 1632, 'steps': 33, 'loss/train': 8.699286460876465}
383
+ 07/25/2024 06:23:11 - INFO - __main__ - Step 35: {'lr': 2.4285714285714285e-05, 'samples': 1680, 'steps': 34, 'loss/train': 8.857367515563965}
384
+ 07/25/2024 06:23:12 - INFO - __main__ - Step 36: {'lr': 2.5e-05, 'samples': 1728, 'steps': 35, 'loss/train': 8.830195426940918}
385
+ 07/25/2024 06:23:12 - INFO - __main__ - Step 37: {'lr': 2.5714285714285714e-05, 'samples': 1776, 'steps': 36, 'loss/train': 8.944982528686523}
386
+ 07/25/2024 06:23:12 - INFO - __main__ - Step 38: {'lr': 2.642857142857143e-05, 'samples': 1824, 'steps': 37, 'loss/train': 8.670278549194336}
387
+ 07/25/2024 06:23:12 - INFO - __main__ - Step 39: {'lr': 2.7142857142857144e-05, 'samples': 1872, 'steps': 38, 'loss/train': 8.710525512695312}
388
+ 07/25/2024 06:23:13 - INFO - __main__ - Step 40: {'lr': 2.7857142857142858e-05, 'samples': 1920, 'steps': 39, 'loss/train': 7.902089595794678}
389
+ 07/25/2024 06:23:13 - INFO - __main__ - Step 41: {'lr': 2.857142857142857e-05, 'samples': 1968, 'steps': 40, 'loss/train': 8.400484085083008}
390
+ 07/25/2024 06:23:13 - INFO - __main__ - Step 42: {'lr': 2.9285714285714288e-05, 'samples': 2016, 'steps': 41, 'loss/train': 8.789310455322266}
391
+ 07/25/2024 06:23:14 - INFO - __main__ - Step 43: {'lr': 3e-05, 'samples': 2064, 'steps': 42, 'loss/train': 8.754344940185547}
392
+ 07/25/2024 06:23:14 - INFO - __main__ - Step 44: {'lr': 3.071428571428572e-05, 'samples': 2112, 'steps': 43, 'loss/train': 8.84192943572998}
393
+ 07/25/2024 06:23:14 - INFO - __main__ - Step 45: {'lr': 3.142857142857143e-05, 'samples': 2160, 'steps': 44, 'loss/train': 8.784793853759766}
394
+ 07/25/2024 06:23:14 - INFO - __main__ - Step 46: {'lr': 3.214285714285714e-05, 'samples': 2208, 'steps': 45, 'loss/train': 8.67403793334961}
395
+ 07/25/2024 06:23:15 - INFO - __main__ - Step 47: {'lr': 3.285714285714286e-05, 'samples': 2256, 'steps': 46, 'loss/train': 8.51427173614502}
396
+ 07/25/2024 06:23:15 - INFO - __main__ - Step 48: {'lr': 3.357142857142857e-05, 'samples': 2304, 'steps': 47, 'loss/train': 8.48193073272705}
397
+ 07/25/2024 06:23:15 - INFO - __main__ - Step 49: {'lr': 3.428571428571429e-05, 'samples': 2352, 'steps': 48, 'loss/train': 8.518038749694824}
398
+ 07/25/2024 06:23:15 - INFO - __main__ - Step 50: {'lr': 3.5000000000000004e-05, 'samples': 2400, 'steps': 49, 'loss/train': 8.63569450378418}
399
+ 07/25/2024 06:23:16 - INFO - __main__ - Evaluating and saving model checkpoint
400
+ 07/25/2024 06:23:16 - DEBUG - datasets.iterable_dataset - dataloader worker#0, ': Starting to iterate over 1/1 shards.
401
+ 07/25/2024 06:23:19 - INFO - __main__ - Step 50: {'loss/eval': 8.551246643066406, 'perplexity': 5173.19970703125}
402
+ 07/25/2024 06:23:20 - INFO - accelerate.accelerator - Saving current state to my_checkpoint
403
+ 07/25/2024 06:23:20 - WARNING - accelerate.utils.other - Removed shared tensor {'lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
404
+ 07/25/2024 06:23:20 - INFO - accelerate.checkpointing - Model weights saved in my_checkpoint/model.safetensors
405
+ 07/25/2024 06:23:21 - INFO - accelerate.checkpointing - Optimizer state saved in my_checkpoint/optimizer.bin
406
+ 07/25/2024 06:23:21 - INFO - accelerate.checkpointing - Sampler state for dataloader 0 saved in my_checkpoint/sampler.bin
407
+ 07/25/2024 06:23:21 - INFO - accelerate.checkpointing - Sampler state for dataloader 1 saved in my_checkpoint/sampler_1.bin
408
+ 07/25/2024 06:23:21 - INFO - accelerate.checkpointing - Gradient scaler state saved in my_checkpoint/scaler.pt
409
+ 07/25/2024 06:23:21 - INFO - accelerate.checkpointing - Random states saved in my_checkpoint/random_states_0.pkl
log/debug_1.log ADDED
File without changes
log/debug_2.log ADDED
File without changes
log/debug_3.log ADDED
File without changes
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3e5bfbaf0ff37c9cbbbf37dfdf848aaee9b060a6a19f3c79d815806112c6c2f1
3
  size 444048000
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7eb13d61f8d3f9cb945838abd62b274c030a665710e11a81e40523fd167676e8
3
  size 444048000
my_checkpoint/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7eb13d61f8d3f9cb945838abd62b274c030a665710e11a81e40523fd167676e8
3
+ size 444048000
my_checkpoint/optimizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e17945a190a7f13238a46e415a9d197e06e630393a465b4eb7395e34ac34cec8
3
+ size 888189882
my_checkpoint/random_states_0.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:836e8625e01d845543882c69cc524c4b56d536fd88b56370ad8514f35d1bb70a
3
+ size 15124
my_checkpoint/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a25e9003157e2a9d5bc24e837058bba42dd027da3fe38540cbd3087114c4852
3
+ size 988
runs/Jul25_06-16-39_lab/1721888199.504291/events.out.tfevents.1721888199.lab.4914.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd5ac20c0a5e7deeacba6057a48d6121bc904a498dae84b1a04c58ab20d0ee0d
3
+ size 1702
runs/Jul25_06-16-39_lab/events.out.tfevents.1721888199.lab.4914.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b19c08ef1e8ddf94fbea9e5ca257dbc6c2f5ed7e1466d2d853425d4228d4fbb
3
+ size 440
runs/Jul25_06-22-39_lab/1721888559.4527462/events.out.tfevents.1721888559.lab.31151.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9c27369a574e442982e72766a30c42813f73348217fb741e56628bf09a00d04
3
+ size 1702
runs/Jul25_06-22-39_lab/events.out.tfevents.1721888559.lab.31151.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:333a3f9007a13708058777a65831eaa49e9d4797f33a45f81f6dfc256407dff1
3
+ size 8983