AlekseyKorshuk
commited on
Commit
•
a5353af
1
Parent(s):
27c76ac
huggingartists
Browse files- README.md +3 -3
- config.json +4 -2
- evaluation.txt +1 -1
- flax_model.msgpack +1 -1
- optimizer.pt +2 -2
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- tokenizer.json +0 -0
- tokenizer_config.json +1 -1
- trainer_state.json +201 -7
- training_args.bin +2 -2
README.md
CHANGED
@@ -45,15 +45,15 @@ from datasets import load_dataset
|
|
45 |
dataset = load_dataset("huggingartists/egor-letov")
|
46 |
```
|
47 |
|
48 |
-
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/
|
49 |
|
50 |
## Training procedure
|
51 |
|
52 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Егор Летов (Egor Letov)'s lyrics.
|
53 |
|
54 |
-
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/
|
55 |
|
56 |
-
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/
|
57 |
|
58 |
## How to use
|
59 |
|
|
|
45 |
dataset = load_dataset("huggingartists/egor-letov")
|
46 |
```
|
47 |
|
48 |
+
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/1turt8ph/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
|
49 |
|
50 |
## Training procedure
|
51 |
|
52 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Егор Летов (Egor Letov)'s lyrics.
|
53 |
|
54 |
+
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/797vcvtd) for full transparency and reproducibility.
|
55 |
|
56 |
+
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/797vcvtd/artifacts) is logged and versioned.
|
57 |
|
58 |
## How to use
|
59 |
|
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_function": "gelu_new",
|
4 |
"architectures": [
|
5 |
"GPT2LMHeadModel"
|
@@ -18,7 +18,9 @@
|
|
18 |
"n_inner": null,
|
19 |
"n_layer": 12,
|
20 |
"n_positions": 1024,
|
|
|
21 |
"resid_pdrop": 0.1,
|
|
|
22 |
"scale_attn_weights": true,
|
23 |
"summary_activation": null,
|
24 |
"summary_first_dropout": 0.1,
|
@@ -35,7 +37,7 @@
|
|
35 |
}
|
36 |
},
|
37 |
"torch_dtype": "float32",
|
38 |
-
"transformers_version": "4.
|
39 |
"use_cache": true,
|
40 |
"vocab_size": 50257
|
41 |
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "egor-letov",
|
3 |
"activation_function": "gelu_new",
|
4 |
"architectures": [
|
5 |
"GPT2LMHeadModel"
|
|
|
18 |
"n_inner": null,
|
19 |
"n_layer": 12,
|
20 |
"n_positions": 1024,
|
21 |
+
"reorder_and_upcast_attn": false,
|
22 |
"resid_pdrop": 0.1,
|
23 |
+
"scale_attn_by_inverse_layer_idx": false,
|
24 |
"scale_attn_weights": true,
|
25 |
"summary_activation": null,
|
26 |
"summary_first_dropout": 0.1,
|
|
|
37 |
}
|
38 |
},
|
39 |
"torch_dtype": "float32",
|
40 |
+
"transformers_version": "4.16.2",
|
41 |
"use_cache": true,
|
42 |
"vocab_size": 50257
|
43 |
}
|
evaluation.txt
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"eval_loss": 1.
|
|
|
1 |
+
{"eval_loss": 1.577836275100708, "eval_runtime": 12.1566, "eval_samples_per_second": 20.647, "eval_steps_per_second": 2.632, "epoch": 2.0}
|
flax_model.msgpack
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 497764120
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:40bb754f5a4e532e1094072560b37d102559f7da5ab2ceca89da9c32a84a7004
|
3 |
size 497764120
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:64e600161fae48afedb06bb4666511c0231bf86944de08c4a4d428ec8abcab03
|
3 |
+
size 995604017
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 510403817
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c71055ec92f1e3ee0ac98e2e1b4a77daead3f0fcf9222865833bcc60ad599f2a
|
3 |
size 510403817
|
rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14503
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3e6a63d522e295e27b8f1e5effa2220fd354b7dd5efe831096984edfac6701c0
|
3 |
size 14503
|
scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bc07e42f7b3eb4494cdbc9bcb8911780143ce3b5a9cba5c3f2005a37846c6c2b
|
3 |
size 623
|
tokenizer.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "
|
|
|
1 |
+
{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "huggingartists/egor-letov", "tokenizer_class": "GPT2Tokenizer"}
|
trainer_state.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"best_metric": 1.
|
3 |
-
"best_model_checkpoint": "output/egor-letov/checkpoint-
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
@@ -206,11 +206,205 @@
|
|
206 |
"eval_samples_per_second": 22.286,
|
207 |
"eval_steps_per_second": 2.872,
|
208 |
"step": 162
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
209 |
}
|
210 |
],
|
211 |
-
"max_steps":
|
212 |
-
"num_train_epochs":
|
213 |
-
"total_flos":
|
214 |
"trial_name": null,
|
215 |
"trial_params": null
|
216 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 1.577836275100708,
|
3 |
+
"best_model_checkpoint": "output/egor-letov/checkpoint-316",
|
4 |
+
"epoch": 2.0,
|
5 |
+
"global_step": 316,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
206 |
"eval_samples_per_second": 22.286,
|
207 |
"eval_steps_per_second": 2.872,
|
208 |
"step": 162
|
209 |
+
},
|
210 |
+
{
|
211 |
+
"epoch": 1.04,
|
212 |
+
"learning_rate": 6.633985999863723e-07,
|
213 |
+
"loss": 1.7473,
|
214 |
+
"step": 165
|
215 |
+
},
|
216 |
+
{
|
217 |
+
"epoch": 1.08,
|
218 |
+
"learning_rate": 1.94348327322787e-06,
|
219 |
+
"loss": 1.676,
|
220 |
+
"step": 170
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 1.11,
|
224 |
+
"learning_rate": 3.881847639061565e-06,
|
225 |
+
"loss": 1.6984,
|
226 |
+
"step": 175
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 1.14,
|
230 |
+
"learning_rate": 6.459348994545111e-06,
|
231 |
+
"loss": 1.7465,
|
232 |
+
"step": 180
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 1.17,
|
236 |
+
"learning_rate": 9.650532712235251e-06,
|
237 |
+
"loss": 1.8561,
|
238 |
+
"step": 185
|
239 |
+
},
|
240 |
+
{
|
241 |
+
"epoch": 1.2,
|
242 |
+
"learning_rate": 1.3423883622425528e-05,
|
243 |
+
"loss": 1.7407,
|
244 |
+
"step": 190
|
245 |
+
},
|
246 |
+
{
|
247 |
+
"epoch": 1.23,
|
248 |
+
"learning_rate": 1.7742137247474782e-05,
|
249 |
+
"loss": 1.7568,
|
250 |
+
"step": 195
|
251 |
+
},
|
252 |
+
{
|
253 |
+
"epoch": 1.27,
|
254 |
+
"learning_rate": 2.256264781457008e-05,
|
255 |
+
"loss": 1.6805,
|
256 |
+
"step": 200
|
257 |
+
},
|
258 |
+
{
|
259 |
+
"epoch": 1.3,
|
260 |
+
"learning_rate": 2.783780941254068e-05,
|
261 |
+
"loss": 1.6899,
|
262 |
+
"step": 205
|
263 |
+
},
|
264 |
+
{
|
265 |
+
"epoch": 1.33,
|
266 |
+
"learning_rate": 3.3515526133505226e-05,
|
267 |
+
"loss": 1.7279,
|
268 |
+
"step": 210
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 1.36,
|
272 |
+
"learning_rate": 3.953972655637492e-05,
|
273 |
+
"loss": 1.7098,
|
274 |
+
"step": 215
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 1.39,
|
278 |
+
"learning_rate": 4.585091749132818e-05,
|
279 |
+
"loss": 1.7613,
|
280 |
+
"step": 220
|
281 |
+
},
|
282 |
+
{
|
283 |
+
"epoch": 1.42,
|
284 |
+
"learning_rate": 5.238677151664272e-05,
|
285 |
+
"loss": 1.635,
|
286 |
+
"step": 225
|
287 |
+
},
|
288 |
+
{
|
289 |
+
"epoch": 1.46,
|
290 |
+
"learning_rate": 5.908274250554988e-05,
|
291 |
+
"loss": 1.7108,
|
292 |
+
"step": 230
|
293 |
+
},
|
294 |
+
{
|
295 |
+
"epoch": 1.49,
|
296 |
+
"learning_rate": 6.587270306435089e-05,
|
297 |
+
"loss": 1.6292,
|
298 |
+
"step": 235
|
299 |
+
},
|
300 |
+
{
|
301 |
+
"epoch": 1.52,
|
302 |
+
"learning_rate": 7.268959758664963e-05,
|
303 |
+
"loss": 1.8167,
|
304 |
+
"step": 240
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 1.55,
|
308 |
+
"learning_rate": 7.946610447433497e-05,
|
309 |
+
"loss": 1.6766,
|
310 |
+
"step": 245
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 1.58,
|
314 |
+
"learning_rate": 8.613530098541655e-05,
|
315 |
+
"loss": 1.5776,
|
316 |
+
"step": 250
|
317 |
+
},
|
318 |
+
{
|
319 |
+
"epoch": 1.61,
|
320 |
+
"learning_rate": 9.263132414287828e-05,
|
321 |
+
"loss": 1.7031,
|
322 |
+
"step": 255
|
323 |
+
},
|
324 |
+
{
|
325 |
+
"epoch": 1.65,
|
326 |
+
"learning_rate": 9.889002117761157e-05,
|
327 |
+
"loss": 1.6967,
|
328 |
+
"step": 260
|
329 |
+
},
|
330 |
+
{
|
331 |
+
"epoch": 1.68,
|
332 |
+
"learning_rate": 0.0001048495830818497,
|
333 |
+
"loss": 1.7448,
|
334 |
+
"step": 265
|
335 |
+
},
|
336 |
+
{
|
337 |
+
"epoch": 1.71,
|
338 |
+
"learning_rate": 0.000110451155016322,
|
339 |
+
"loss": 1.677,
|
340 |
+
"step": 270
|
341 |
+
},
|
342 |
+
{
|
343 |
+
"epoch": 1.74,
|
344 |
+
"learning_rate": 0.00011563941754292792,
|
345 |
+
"loss": 1.7431,
|
346 |
+
"step": 275
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 1.77,
|
350 |
+
"learning_rate": 0.00012036313294285455,
|
351 |
+
"loss": 1.7339,
|
352 |
+
"step": 280
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 1.8,
|
356 |
+
"learning_rate": 0.00012457565122486245,
|
357 |
+
"loss": 1.7054,
|
358 |
+
"step": 285
|
359 |
+
},
|
360 |
+
{
|
361 |
+
"epoch": 1.84,
|
362 |
+
"learning_rate": 0.00012823537082655537,
|
363 |
+
"loss": 1.6155,
|
364 |
+
"step": 290
|
365 |
+
},
|
366 |
+
{
|
367 |
+
"epoch": 1.87,
|
368 |
+
"learning_rate": 0.00013130614945888676,
|
369 |
+
"loss": 1.7792,
|
370 |
+
"step": 295
|
371 |
+
},
|
372 |
+
{
|
373 |
+
"epoch": 1.9,
|
374 |
+
"learning_rate": 0.0001337576610365259,
|
375 |
+
"loss": 1.8149,
|
376 |
+
"step": 300
|
377 |
+
},
|
378 |
+
{
|
379 |
+
"epoch": 1.93,
|
380 |
+
"learning_rate": 0.0001355656951691469,
|
381 |
+
"loss": 1.686,
|
382 |
+
"step": 305
|
383 |
+
},
|
384 |
+
{
|
385 |
+
"epoch": 1.96,
|
386 |
+
"learning_rate": 0.00013671239625595536,
|
387 |
+
"loss": 1.6878,
|
388 |
+
"step": 310
|
389 |
+
},
|
390 |
+
{
|
391 |
+
"epoch": 1.99,
|
392 |
+
"learning_rate": 0.0001371864398222265,
|
393 |
+
"loss": 1.6786,
|
394 |
+
"step": 315
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 2.0,
|
398 |
+
"eval_loss": 1.577836275100708,
|
399 |
+
"eval_runtime": 12.0961,
|
400 |
+
"eval_samples_per_second": 20.75,
|
401 |
+
"eval_steps_per_second": 2.645,
|
402 |
+
"step": 316
|
403 |
}
|
404 |
],
|
405 |
+
"max_steps": 316,
|
406 |
+
"num_train_epochs": 2,
|
407 |
+
"total_flos": 329097314304000.0,
|
408 |
"trial_name": null,
|
409 |
"trial_params": null
|
410 |
}
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e5a982dd4f44bc59417f7e482b5d0cf9f6714929b958f9dd7f250c26a4889aca
|
3 |
+
size 3055
|