AlekseyKorshuk commited on
Commit
cb31c93
1 Parent(s): 798bc92

huggingartists

Browse files
README.md CHANGED
@@ -14,7 +14,7 @@ widget:
14
  <div class="inline-flex flex-col" style="line-height: 1.5;">
15
  <div class="flex">
16
  <div
17
- style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url(&#39;https://images.genius.com/5cad1fbda7135f40c49ea35ad60ba4f5.215x215x1.png&#39;)">
18
  </div>
19
  </div>
20
  <div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
@@ -45,15 +45,15 @@ from datasets import load_dataset
45
  dataset = load_dataset("huggingartists/madonna")
46
  ```
47
 
48
- [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/2eqc1vv3/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 Madonna's lyrics.
53
 
54
- Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/ppm2eyoj) for full transparency and reproducibility.
55
 
56
- At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/ppm2eyoj/artifacts) is logged and versioned.
57
 
58
  ## How to use
59
 
 
14
  <div class="inline-flex flex-col" style="line-height: 1.5;">
15
  <div class="flex">
16
  <div
17
+ style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url(&#39;https://images.genius.com/e4051ba629f57622a2a35571e5676824.720x720x1.jpg&#39;)">
18
  </div>
19
  </div>
20
  <div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
 
45
  dataset = load_dataset("huggingartists/madonna")
46
  ```
47
 
48
+ [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/1zgkgxrp/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 Madonna's lyrics.
53
 
54
+ Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1nejy753) for full transparency and reproducibility.
55
 
56
+ At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1nejy753/artifacts) is logged and versioned.
57
 
58
  ## How to use
59
 
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "gpt2",
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": "4.10.2",
39
  "use_cache": true,
40
  "vocab_size": 50257
41
  }
 
1
  {
2
+ "_name_or_path": "madonna",
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": "4.15.0",
41
  "use_cache": true,
42
  "vocab_size": 50257
43
  }
evaluation.txt CHANGED
@@ -1 +1 @@
1
- {"eval_loss": 3.1591076850891113, "eval_runtime": 3.3487, "eval_samples_per_second": 21.202, "eval_steps_per_second": 2.688, "epoch": 1.0}
 
1
+ {"eval_loss": 1.5412778854370117, "eval_runtime": 17.7742, "eval_samples_per_second": 17.272, "eval_steps_per_second": 2.194, "epoch": 2.0}
flax_model.msgpack CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b9b6a9dde4f77fe7e5286411ef04c90642d3d5632397b8526fcc446a7864b4f1
3
  size 497764120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25566d27aa0a638f5602254e69031264732abeaa33b25755bd5b7e9febdce8f7
3
  size 497764120
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ceae229ca904bed5d1bb6cd9932ed1f1ce4b1f959602365ba45f365e7ae157b8
3
- size 995603825
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dc55df2228257ecd9fbf03f2979f7659de46424b72ffa514ec925c65c4ceaf50
3
+ size 995604017
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a28db54f1ddd728c41eb92a39efd92a5721e39626fd0317625976c4d82ef50fb
3
  size 510403817
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cffe2698db26bffad1da8bea41d71775f16f1b35e022d24c8c80b1ca1297a533
3
  size 510403817
rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9d2957881bb5b55492f8fff01b392d863d09b0e66192bf730f50b9b2472170d6
3
  size 14503
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0706774df0979b928412046725f3584988207df65c21b575ba45213c1e46e201
3
  size 14503
scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:33ebdf640e584fb96b06bd917ad6a00182e9ef4871b5e81f588509ccfb511410
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73d66e37400481ce71994edfed806089e04a5cb8f448f68f1ec267b591d1dbe6
3
  size 623
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": "huggingartists/zemfira", "tokenizer_class": "GPT2Tokenizer"}
 
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/madonna", "tokenizer_class": "GPT2Tokenizer"}
trainer_state.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "best_metric": 1.924851655960083,
3
- "best_model_checkpoint": "output/madonna/checkpoint-222",
4
- "epoch": 1.0,
5
- "global_step": 222,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
@@ -278,11 +278,297 @@
278
  "eval_samples_per_second": 22.672,
279
  "eval_steps_per_second": 2.887,
280
  "step": 222
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
281
  }
282
  ],
283
- "max_steps": 222,
284
- "num_train_epochs": 1,
285
- "total_flos": 232027324416000.0,
286
  "trial_name": null,
287
  "trial_params": null
288
  }
 
1
  {
2
+ "best_metric": 1.5412778854370117,
3
+ "best_model_checkpoint": "output/madonna/checkpoint-448",
4
+ "epoch": 2.0,
5
+ "global_step": 448,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
 
278
  "eval_samples_per_second": 22.672,
279
  "eval_steps_per_second": 2.887,
280
  "step": 222
281
+ },
282
+ {
283
+ "epoch": 1.0,
284
+ "eval_loss": 1.7440342903137207,
285
+ "eval_runtime": 17.6835,
286
+ "eval_samples_per_second": 17.361,
287
+ "eval_steps_per_second": 2.205,
288
+ "step": 224
289
+ },
290
+ {
291
+ "epoch": 1.0,
292
+ "learning_rate": 6.746689293153874e-09,
293
+ "loss": 2.1578,
294
+ "step": 225
295
+ },
296
+ {
297
+ "epoch": 1.03,
298
+ "learning_rate": 2.427415035313232e-07,
299
+ "loss": 1.867,
300
+ "step": 230
301
+ },
302
+ {
303
+ "epoch": 1.05,
304
+ "learning_rate": 8.147449052190903e-07,
305
+ "loss": 1.7004,
306
+ "step": 235
307
+ },
308
+ {
309
+ "epoch": 1.07,
310
+ "learning_rate": 1.7199452243268996e-06,
311
+ "loss": 1.8748,
312
+ "step": 240
313
+ },
314
+ {
315
+ "epoch": 1.09,
316
+ "learning_rate": 2.9538929687704672e-06,
317
+ "loss": 1.7175,
318
+ "step": 245
319
+ },
320
+ {
321
+ "epoch": 1.12,
322
+ "learning_rate": 4.510522695790453e-06,
323
+ "loss": 1.9134,
324
+ "step": 250
325
+ },
326
+ {
327
+ "epoch": 1.14,
328
+ "learning_rate": 6.382182826501689e-06,
329
+ "loss": 1.8779,
330
+ "step": 255
331
+ },
332
+ {
333
+ "epoch": 1.16,
334
+ "learning_rate": 8.559673257059505e-06,
335
+ "loss": 1.9084,
336
+ "step": 260
337
+ },
338
+ {
339
+ "epoch": 1.18,
340
+ "learning_rate": 1.1032290581565944e-05,
341
+ "loss": 1.7949,
342
+ "step": 265
343
+ },
344
+ {
345
+ "epoch": 1.21,
346
+ "learning_rate": 1.3787880704424101e-05,
347
+ "loss": 1.7769,
348
+ "step": 270
349
+ },
350
+ {
351
+ "epoch": 1.23,
352
+ "learning_rate": 1.681289858352526e-05,
353
+ "loss": 1.724,
354
+ "step": 275
355
+ },
356
+ {
357
+ "epoch": 1.25,
358
+ "learning_rate": 2.009247481060283e-05,
359
+ "loss": 1.8311,
360
+ "step": 280
361
+ },
362
+ {
363
+ "epoch": 1.27,
364
+ "learning_rate": 2.3610488701478854e-05,
365
+ "loss": 1.7348,
366
+ "step": 285
367
+ },
368
+ {
369
+ "epoch": 1.29,
370
+ "learning_rate": 2.734964753692934e-05,
371
+ "loss": 1.768,
372
+ "step": 290
373
+ },
374
+ {
375
+ "epoch": 1.32,
376
+ "learning_rate": 3.129157156466198e-05,
377
+ "loss": 1.7488,
378
+ "step": 295
379
+ },
380
+ {
381
+ "epoch": 1.34,
382
+ "learning_rate": 3.541688434458052e-05,
383
+ "loss": 1.9316,
384
+ "step": 300
385
+ },
386
+ {
387
+ "epoch": 1.36,
388
+ "learning_rate": 3.970530799324656e-05,
389
+ "loss": 1.7607,
390
+ "step": 305
391
+ },
392
+ {
393
+ "epoch": 1.38,
394
+ "learning_rate": 4.413576285936533e-05,
395
+ "loss": 1.6681,
396
+ "step": 310
397
+ },
398
+ {
399
+ "epoch": 1.41,
400
+ "learning_rate": 4.8686471140343875e-05,
401
+ "loss": 2.065,
402
+ "step": 315
403
+ },
404
+ {
405
+ "epoch": 1.43,
406
+ "learning_rate": 5.333506393059682e-05,
407
+ "loss": 1.848,
408
+ "step": 320
409
+ },
410
+ {
411
+ "epoch": 1.45,
412
+ "learning_rate": 5.805869117540521e-05,
413
+ "loss": 1.8802,
414
+ "step": 325
415
+ },
416
+ {
417
+ "epoch": 1.47,
418
+ "learning_rate": 6.283413398985365e-05,
419
+ "loss": 1.7551,
420
+ "step": 330
421
+ },
422
+ {
423
+ "epoch": 1.5,
424
+ "learning_rate": 6.763791879074358e-05,
425
+ "loss": 1.7544,
426
+ "step": 335
427
+ },
428
+ {
429
+ "epoch": 1.52,
430
+ "learning_rate": 7.244643268047132e-05,
431
+ "loss": 1.793,
432
+ "step": 340
433
+ },
434
+ {
435
+ "epoch": 1.54,
436
+ "learning_rate": 7.723603951570171e-05,
437
+ "loss": 1.7917,
438
+ "step": 345
439
+ },
440
+ {
441
+ "epoch": 1.56,
442
+ "learning_rate": 8.19831960903064e-05,
443
+ "loss": 1.63,
444
+ "step": 350
445
+ },
446
+ {
447
+ "epoch": 1.58,
448
+ "learning_rate": 8.666456786147141e-05,
449
+ "loss": 2.0457,
450
+ "step": 355
451
+ },
452
+ {
453
+ "epoch": 1.61,
454
+ "learning_rate": 9.125714365012444e-05,
455
+ "loss": 2.0322,
456
+ "step": 360
457
+ },
458
+ {
459
+ "epoch": 1.63,
460
+ "learning_rate": 9.5738348751875e-05,
461
+ "loss": 1.687,
462
+ "step": 365
463
+ },
464
+ {
465
+ "epoch": 1.65,
466
+ "learning_rate": 0.00010008615590247305,
467
+ "loss": 1.9025,
468
+ "step": 370
469
+ },
470
+ {
471
+ "epoch": 1.67,
472
+ "learning_rate": 0.00010427919355233763,
473
+ "loss": 1.7503,
474
+ "step": 375
475
+ },
476
+ {
477
+ "epoch": 1.7,
478
+ "learning_rate": 0.00010829685091793463,
479
+ "loss": 1.7749,
480
+ "step": 380
481
+ },
482
+ {
483
+ "epoch": 1.72,
484
+ "learning_rate": 0.00011211937929362609,
485
+ "loss": 1.6852,
486
+ "step": 385
487
+ },
488
+ {
489
+ "epoch": 1.74,
490
+ "learning_rate": 0.00011572798912599424,
491
+ "loss": 1.8013,
492
+ "step": 390
493
+ },
494
+ {
495
+ "epoch": 1.76,
496
+ "learning_rate": 0.00011910494237347491,
497
+ "loss": 1.5069,
498
+ "step": 395
499
+ },
500
+ {
501
+ "epoch": 1.79,
502
+ "learning_rate": 0.00012223363969730684,
503
+ "loss": 1.5859,
504
+ "step": 400
505
+ },
506
+ {
507
+ "epoch": 1.81,
508
+ "learning_rate": 0.00012509870205521334,
509
+ "loss": 1.5894,
510
+ "step": 405
511
+ },
512
+ {
513
+ "epoch": 1.83,
514
+ "learning_rate": 0.00012768604629674322,
515
+ "loss": 1.9085,
516
+ "step": 410
517
+ },
518
+ {
519
+ "epoch": 1.85,
520
+ "learning_rate": 0.00012998295438868449,
521
+ "loss": 1.8982,
522
+ "step": 415
523
+ },
524
+ {
525
+ "epoch": 1.88,
526
+ "learning_rate": 0.00013197813593027427,
527
+ "loss": 1.9507,
528
+ "step": 420
529
+ },
530
+ {
531
+ "epoch": 1.9,
532
+ "learning_rate": 0.00013366178365091335,
533
+ "loss": 1.7504,
534
+ "step": 425
535
+ },
536
+ {
537
+ "epoch": 1.92,
538
+ "learning_rate": 0.00013502562161758715,
539
+ "loss": 1.7843,
540
+ "step": 430
541
+ },
542
+ {
543
+ "epoch": 1.94,
544
+ "learning_rate": 0.00013606294591503093,
545
+ "loss": 1.6083,
546
+ "step": 435
547
+ },
548
+ {
549
+ "epoch": 1.96,
550
+ "learning_rate": 0.00013676865759867644,
551
+ "loss": 1.5864,
552
+ "step": 440
553
+ },
554
+ {
555
+ "epoch": 1.99,
556
+ "learning_rate": 0.00013713928775840084,
557
+ "loss": 1.8849,
558
+ "step": 445
559
+ },
560
+ {
561
+ "epoch": 2.0,
562
+ "eval_loss": 1.5412778854370117,
563
+ "eval_runtime": 17.7637,
564
+ "eval_samples_per_second": 17.282,
565
+ "eval_steps_per_second": 2.195,
566
+ "step": 448
567
  }
568
  ],
569
+ "max_steps": 448,
570
+ "num_train_epochs": 2,
571
+ "total_flos": 467712737280000.0,
572
  "trial_name": null,
573
  "trial_params": null
574
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:83c9e6cfb63c365b4c49a065df2aef2c18c8bd82cedbf8864df569b4417ff289
3
- size 2671
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eaf00d98777c40d1ee6ae43b9a54114156e44aec0020e9e8e22c522be0398be6
3
+ size 2991