Training in progress, step 500
Browse files- all_results.json +8 -0
- config.json +2 -2
- pytorch_model.bin +2 -2
- runs/Sep08_10-32-10_aware/1662626035.8880892/events.out.tfevents.1662626035.aware.352107.1 +3 -0
- runs/Sep08_10-32-10_aware/events.out.tfevents.1662626035.aware.352107.0 +3 -0
- runs/Sep08_10-49-36_aware/1662627005.7353008/events.out.tfevents.1662627005.aware.816778.1 +3 -0
- runs/Sep08_10-49-36_aware/events.out.tfevents.1662627005.aware.816778.0 +3 -0
- runs/Sep08_10-52-49_aware/1662627189.2480614/events.out.tfevents.1662627189.aware.910075.1 +3 -0
- runs/Sep08_10-52-49_aware/events.out.tfevents.1662627189.aware.910075.0 +3 -0
- runs/Sep08_10-57-41_aware/1662627483.804511/events.out.tfevents.1662627483.aware.1065203.1 +3 -0
- runs/Sep08_10-57-41_aware/events.out.tfevents.1662627483.aware.1065203.0 +3 -0
- runs/Sep08_11-06-03_aware/1662627994.1538968/events.out.tfevents.1662627994.aware.1355581.1 +3 -0
- runs/Sep08_11-06-03_aware/events.out.tfevents.1662627994.aware.1355581.0 +3 -0
- runs/Sep08_11-21-06_aware/1662628891.6121929/events.out.tfevents.1662628891.aware.1820222.1 +3 -0
- runs/Sep08_11-21-06_aware/events.out.tfevents.1662628891.aware.1820222.0 +3 -0
- runs/Sep08_11-29-02_aware/1662629370.4165187/events.out.tfevents.1662629370.aware.2018942.1 +3 -0
- runs/Sep08_11-29-02_aware/events.out.tfevents.1662629370.aware.2018942.0 +3 -0
- runs/Sep08_11-33-00_aware/1662629603.2161145/events.out.tfevents.1662629603.aware.2084717.1 +3 -0
- runs/Sep08_11-33-00_aware/events.out.tfevents.1662629603.aware.2084717.0 +3 -0
- special_tokens_map.json +252 -0
- train_results.json +8 -0
- trainer_state.json +0 -0
- training_args.bin +1 -1
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 3.1717406544744464,
|
| 4 |
+
"train_runtime": 43485.0285,
|
| 5 |
+
"train_samples": 458274,
|
| 6 |
+
"train_samples_per_second": 31.616,
|
| 7 |
+
"train_steps_per_second": 0.035
|
| 8 |
+
}
|
config.json
CHANGED
|
@@ -29,14 +29,14 @@
|
|
| 29 |
"initializer_range": 0.02,
|
| 30 |
"input_channels": 1,
|
| 31 |
"input_feat_per_channel": 80,
|
| 32 |
-
"intermediate_size":
|
| 33 |
"layer_norm_eps": 1e-05,
|
| 34 |
"layerdrop": 0.0,
|
| 35 |
"mask_feature_length": 10,
|
| 36 |
"mask_feature_prob": 0.0,
|
| 37 |
"mask_time_length": 10,
|
| 38 |
"mask_time_prob": 0.0,
|
| 39 |
-
"max_position_embeddings":
|
| 40 |
"model_type": "mctct",
|
| 41 |
"num_attention_heads": 4,
|
| 42 |
"num_conv_layers": 1,
|
|
|
|
| 29 |
"initializer_range": 0.02,
|
| 30 |
"input_channels": 1,
|
| 31 |
"input_feat_per_channel": 80,
|
| 32 |
+
"intermediate_size": 6144,
|
| 33 |
"layer_norm_eps": 1e-05,
|
| 34 |
"layerdrop": 0.0,
|
| 35 |
"mask_feature_length": 10,
|
| 36 |
"mask_feature_prob": 0.0,
|
| 37 |
"mask_time_length": 10,
|
| 38 |
"mask_time_prob": 0.0,
|
| 39 |
+
"max_position_embeddings": 920,
|
| 40 |
"model_type": "mctct",
|
| 41 |
"num_attention_heads": 4,
|
| 42 |
"num_conv_layers": 1,
|
pytorch_model.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:dd21e157c201d14e1b8a1533df232cd6353a2b53e31d397ef523f076c97d0481
|
| 3 |
+
size 1052199791
|
runs/Sep08_10-32-10_aware/1662626035.8880892/events.out.tfevents.1662626035.aware.352107.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be0cc87415bfc53f58e3cd7c299f2d97d4f2feb15edf4331b987954b121206d1
|
| 3 |
+
size 5667
|
runs/Sep08_10-32-10_aware/events.out.tfevents.1662626035.aware.352107.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6358f587db7599e7655af373a18552785c32a15b79cd48b1c1053b587e2b6697
|
| 3 |
+
size 10969
|
runs/Sep08_10-49-36_aware/1662627005.7353008/events.out.tfevents.1662627005.aware.816778.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62459da5f913827fadcb33359abc730ffce7ff20543d41495913a599972201b1
|
| 3 |
+
size 5667
|
runs/Sep08_10-49-36_aware/events.out.tfevents.1662627005.aware.816778.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0e9643061d82b879960d458ec6f93ae80c69b6cca04fab46144de4a72388251e
|
| 3 |
+
size 4964
|
runs/Sep08_10-52-49_aware/1662627189.2480614/events.out.tfevents.1662627189.aware.910075.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:500a5a724ad384a4a21c4227da7c082fec3f2058037bcdc176dfe4146b6031d3
|
| 3 |
+
size 5667
|
runs/Sep08_10-52-49_aware/events.out.tfevents.1662627189.aware.910075.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84763579ab539d1cdf56d001bd3066ee49eb0f50b8175497a6dc0813f95d6311
|
| 3 |
+
size 6966
|
runs/Sep08_10-57-41_aware/1662627483.804511/events.out.tfevents.1662627483.aware.1065203.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c16c62d18f3f945d64a6e03f214417ccda2a5d87768331af2f18b7360d479511
|
| 3 |
+
size 5667
|
runs/Sep08_10-57-41_aware/events.out.tfevents.1662627483.aware.1065203.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d667bf3f51b0365f111c78b9abc57e52dbc7464a0d09419c13c6bae2ec3d01a3
|
| 3 |
+
size 8044
|
runs/Sep08_11-06-03_aware/1662627994.1538968/events.out.tfevents.1662627994.aware.1355581.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07c1beef46201e78d53c1e3ef5b20d6e20e565e03e3d45e457acb1f149fbfb14
|
| 3 |
+
size 5667
|
runs/Sep08_11-06-03_aware/events.out.tfevents.1662627994.aware.1355581.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7eaf72e02731895fd1480385e8fefc0d1a57408155dbe548c0c669bd16726d62
|
| 3 |
+
size 11278
|
runs/Sep08_11-21-06_aware/1662628891.6121929/events.out.tfevents.1662628891.aware.1820222.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f90585f1f19811acbdf2eb9fedc9a08ce9a1a549fa6cee617794dbd967a4edec
|
| 3 |
+
size 5667
|
runs/Sep08_11-21-06_aware/events.out.tfevents.1662628891.aware.1820222.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ab392cab6680788e633fe2209da47e9d01efea45283c145d8dbb198b3e3528f
|
| 3 |
+
size 8506
|
runs/Sep08_11-29-02_aware/1662629370.4165187/events.out.tfevents.1662629370.aware.2018942.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ba8250fb125bad932707125e2c6103d024224b92bab0ddb14fbc0ee083de3e6
|
| 3 |
+
size 5667
|
runs/Sep08_11-29-02_aware/events.out.tfevents.1662629370.aware.2018942.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:13a880179d834467efb3292c192bed0409add8c05bb7b6c459e5c9194f288d6a
|
| 3 |
+
size 7120
|
runs/Sep08_11-33-00_aware/1662629603.2161145/events.out.tfevents.1662629603.aware.2084717.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94a1421ebec7d1593426f8cdda371988b6f5fe53b6328f8f8cb0b7c2cd18b78f
|
| 3 |
+
size 5667
|
runs/Sep08_11-33-00_aware/events.out.tfevents.1662629603.aware.2084717.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ff6c7a0cfcf3fedbf594ae82d61cff4666aabdbb7aa05925fbdfde916376327
|
| 3 |
+
size 82467
|
special_tokens_map.json
CHANGED
|
@@ -497,6 +497,258 @@
|
|
| 497 |
"rstrip": false,
|
| 498 |
"single_word": false
|
| 499 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 500 |
{
|
| 501 |
"content": "</s>",
|
| 502 |
"lstrip": false,
|
|
|
|
| 497 |
"rstrip": false,
|
| 498 |
"single_word": false
|
| 499 |
},
|
| 500 |
+
{
|
| 501 |
+
"content": "</s>",
|
| 502 |
+
"lstrip": false,
|
| 503 |
+
"normalized": true,
|
| 504 |
+
"rstrip": false,
|
| 505 |
+
"single_word": false
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"content": "<s>",
|
| 509 |
+
"lstrip": false,
|
| 510 |
+
"normalized": true,
|
| 511 |
+
"rstrip": false,
|
| 512 |
+
"single_word": false
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"content": "</s>",
|
| 516 |
+
"lstrip": false,
|
| 517 |
+
"normalized": true,
|
| 518 |
+
"rstrip": false,
|
| 519 |
+
"single_word": false
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"content": "<s>",
|
| 523 |
+
"lstrip": false,
|
| 524 |
+
"normalized": true,
|
| 525 |
+
"rstrip": false,
|
| 526 |
+
"single_word": false
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"content": "</s>",
|
| 530 |
+
"lstrip": false,
|
| 531 |
+
"normalized": true,
|
| 532 |
+
"rstrip": false,
|
| 533 |
+
"single_word": false
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"content": "<s>",
|
| 537 |
+
"lstrip": false,
|
| 538 |
+
"normalized": true,
|
| 539 |
+
"rstrip": false,
|
| 540 |
+
"single_word": false
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"content": "</s>",
|
| 544 |
+
"lstrip": false,
|
| 545 |
+
"normalized": true,
|
| 546 |
+
"rstrip": false,
|
| 547 |
+
"single_word": false
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"content": "<s>",
|
| 551 |
+
"lstrip": false,
|
| 552 |
+
"normalized": true,
|
| 553 |
+
"rstrip": false,
|
| 554 |
+
"single_word": false
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"content": "</s>",
|
| 558 |
+
"lstrip": false,
|
| 559 |
+
"normalized": true,
|
| 560 |
+
"rstrip": false,
|
| 561 |
+
"single_word": false
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"content": "<s>",
|
| 565 |
+
"lstrip": false,
|
| 566 |
+
"normalized": true,
|
| 567 |
+
"rstrip": false,
|
| 568 |
+
"single_word": false
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"content": "</s>",
|
| 572 |
+
"lstrip": false,
|
| 573 |
+
"normalized": true,
|
| 574 |
+
"rstrip": false,
|
| 575 |
+
"single_word": false
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"content": "<s>",
|
| 579 |
+
"lstrip": false,
|
| 580 |
+
"normalized": true,
|
| 581 |
+
"rstrip": false,
|
| 582 |
+
"single_word": false
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"content": "</s>",
|
| 586 |
+
"lstrip": false,
|
| 587 |
+
"normalized": true,
|
| 588 |
+
"rstrip": false,
|
| 589 |
+
"single_word": false
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"content": "<s>",
|
| 593 |
+
"lstrip": false,
|
| 594 |
+
"normalized": true,
|
| 595 |
+
"rstrip": false,
|
| 596 |
+
"single_word": false
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"content": "</s>",
|
| 600 |
+
"lstrip": false,
|
| 601 |
+
"normalized": true,
|
| 602 |
+
"rstrip": false,
|
| 603 |
+
"single_word": false
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"content": "<s>",
|
| 607 |
+
"lstrip": false,
|
| 608 |
+
"normalized": true,
|
| 609 |
+
"rstrip": false,
|
| 610 |
+
"single_word": false
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"content": "</s>",
|
| 614 |
+
"lstrip": false,
|
| 615 |
+
"normalized": true,
|
| 616 |
+
"rstrip": false,
|
| 617 |
+
"single_word": false
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"content": "<s>",
|
| 621 |
+
"lstrip": false,
|
| 622 |
+
"normalized": true,
|
| 623 |
+
"rstrip": false,
|
| 624 |
+
"single_word": false
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"content": "</s>",
|
| 628 |
+
"lstrip": false,
|
| 629 |
+
"normalized": true,
|
| 630 |
+
"rstrip": false,
|
| 631 |
+
"single_word": false
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"content": "<s>",
|
| 635 |
+
"lstrip": false,
|
| 636 |
+
"normalized": true,
|
| 637 |
+
"rstrip": false,
|
| 638 |
+
"single_word": false
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"content": "</s>",
|
| 642 |
+
"lstrip": false,
|
| 643 |
+
"normalized": true,
|
| 644 |
+
"rstrip": false,
|
| 645 |
+
"single_word": false
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"content": "<s>",
|
| 649 |
+
"lstrip": false,
|
| 650 |
+
"normalized": true,
|
| 651 |
+
"rstrip": false,
|
| 652 |
+
"single_word": false
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"content": "</s>",
|
| 656 |
+
"lstrip": false,
|
| 657 |
+
"normalized": true,
|
| 658 |
+
"rstrip": false,
|
| 659 |
+
"single_word": false
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"content": "<s>",
|
| 663 |
+
"lstrip": false,
|
| 664 |
+
"normalized": true,
|
| 665 |
+
"rstrip": false,
|
| 666 |
+
"single_word": false
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"content": "</s>",
|
| 670 |
+
"lstrip": false,
|
| 671 |
+
"normalized": true,
|
| 672 |
+
"rstrip": false,
|
| 673 |
+
"single_word": false
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"content": "<s>",
|
| 677 |
+
"lstrip": false,
|
| 678 |
+
"normalized": true,
|
| 679 |
+
"rstrip": false,
|
| 680 |
+
"single_word": false
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"content": "</s>",
|
| 684 |
+
"lstrip": false,
|
| 685 |
+
"normalized": true,
|
| 686 |
+
"rstrip": false,
|
| 687 |
+
"single_word": false
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"content": "<s>",
|
| 691 |
+
"lstrip": false,
|
| 692 |
+
"normalized": true,
|
| 693 |
+
"rstrip": false,
|
| 694 |
+
"single_word": false
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"content": "</s>",
|
| 698 |
+
"lstrip": false,
|
| 699 |
+
"normalized": true,
|
| 700 |
+
"rstrip": false,
|
| 701 |
+
"single_word": false
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"content": "<s>",
|
| 705 |
+
"lstrip": false,
|
| 706 |
+
"normalized": true,
|
| 707 |
+
"rstrip": false,
|
| 708 |
+
"single_word": false
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"content": "</s>",
|
| 712 |
+
"lstrip": false,
|
| 713 |
+
"normalized": true,
|
| 714 |
+
"rstrip": false,
|
| 715 |
+
"single_word": false
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"content": "<s>",
|
| 719 |
+
"lstrip": false,
|
| 720 |
+
"normalized": true,
|
| 721 |
+
"rstrip": false,
|
| 722 |
+
"single_word": false
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"content": "</s>",
|
| 726 |
+
"lstrip": false,
|
| 727 |
+
"normalized": true,
|
| 728 |
+
"rstrip": false,
|
| 729 |
+
"single_word": false
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"content": "<s>",
|
| 733 |
+
"lstrip": false,
|
| 734 |
+
"normalized": true,
|
| 735 |
+
"rstrip": false,
|
| 736 |
+
"single_word": false
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"content": "</s>",
|
| 740 |
+
"lstrip": false,
|
| 741 |
+
"normalized": true,
|
| 742 |
+
"rstrip": false,
|
| 743 |
+
"single_word": false
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"content": "<s>",
|
| 747 |
+
"lstrip": false,
|
| 748 |
+
"normalized": true,
|
| 749 |
+
"rstrip": false,
|
| 750 |
+
"single_word": false
|
| 751 |
+
},
|
| 752 |
{
|
| 753 |
"content": "</s>",
|
| 754 |
"lstrip": false,
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 3.1717406544744464,
|
| 4 |
+
"train_runtime": 43485.0285,
|
| 5 |
+
"train_samples": 458274,
|
| 6 |
+
"train_samples_per_second": 31.616,
|
| 7 |
+
"train_steps_per_second": 0.035
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b37d36c047b8f84056f5b37e32815b594e8ef0c88a3de8dffc24ada0139d1968
|
| 3 |
size 3503
|