Hugging Face's logo

ajtorek
/
Electra-Test-Model-BabyLM

TensorBoard
Safetensors
electra
Model card Files
xet
Metrics Community
ajtorek commited on
Commit
d66604e
·
verified ·
1 Parent(s): 85ba5b9

Training in progress, step 10000

Browse files
Files changed (43) hide show
  1. config.json +30 -0
  2. model.safetensors +3 -0
  3. runs/Nov11_20-53-25_gpu113/events.out.tfevents.1762919606.gpu113.4215.0 +3 -0
  4. runs/Nov11_20-53-25_gpu113/events.out.tfevents.1762919606.gpu113.4216.0 +3 -0
  5. runs/Nov11_20-53-26_gpu113/events.out.tfevents.1762919606.gpu113.4217.0 +3 -0
  6. runs/Nov11_20-53-26_gpu113/events.out.tfevents.1762919606.gpu113.4218.0 +3 -0
  7. runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5496.0 +3 -0
  8. runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5497.0 +3 -0
  9. runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5498.0 +3 -0
  10. runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5499.0 +3 -0
  11. runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6519.0 +3 -0
  12. runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6520.0 +3 -0
  13. runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6521.0 +3 -0
  14. runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6522.0 +3 -0
  15. runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7427.0 +3 -0
  16. runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7428.0 +3 -0
  17. runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7429.0 +3 -0
  18. runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7430.0 +3 -0
  19. runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8533.0 +3 -0
  20. runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8534.0 +3 -0
  21. runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8535.0 +3 -0
  22. runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8536.0 +3 -0
  23. runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9634.0 +3 -0
  24. runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9635.0 +3 -0
  25. runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9636.0 +3 -0
  26. runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9637.0 +3 -0
  27. runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11623.0 +3 -0
  28. runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11624.0 +3 -0
  29. runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11625.0 +3 -0
  30. runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11626.0 +3 -0
  31. runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924909.gpu111.56195.0 +3 -0
  32. runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56192.0 +3 -0
  33. runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56193.0 +3 -0
  34. runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56194.0 +3 -0
  35. runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44607.0 +3 -0
  36. runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44608.0 +3 -0
  37. runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44609.0 +3 -0
  38. runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44610.0 +3 -0
  39. runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52598.0 +3 -0
  40. runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52599.0 +3 -0
  41. runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52600.0 +3 -0
  42. runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52601.0 +3 -0
  43. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/electra-small-discriminator",
3
+ "architectures": [
4
+ "ElectraForMaskedLM"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "embedding_size": 128,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 256,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 1024,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "electra",
17
+ "num_attention_heads": 4,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 0,
20
+ "position_embedding_type": "absolute",
21
+ "summary_activation": "gelu",
22
+ "summary_last_dropout": 0.1,
23
+ "summary_type": "first",
24
+ "summary_use_proj": true,
25
+ "torch_dtype": "float32",
26
+ "transformers_version": "4.48.3",
27
+ "type_vocab_size": 2,
28
+ "use_cache": true,
29
+ "vocab_size": 30522
30
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74803e113c8762c4d644457d3172dfbff4db0d16b48514667f77fe00fd6dbdc7
3
+ size 54210824
runs/Nov11_20-53-25_gpu113/events.out.tfevents.1762919606.gpu113.4215.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3622a73ba0687125bc6c6416fba371444c4aade837c5f2877cf3162ee7f9d015
3
+ size 5220
runs/Nov11_20-53-25_gpu113/events.out.tfevents.1762919606.gpu113.4216.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10258ee37ec9eedaf805464d574898b090528225cd593f5bb3c210238da4683e
3
+ size 5220
runs/Nov11_20-53-26_gpu113/events.out.tfevents.1762919606.gpu113.4217.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55943f1c2fb16fb6a126d16ef4b49f68c5f8f2801f527457d97da54548ebc1ba
3
+ size 5220
runs/Nov11_20-53-26_gpu113/events.out.tfevents.1762919606.gpu113.4218.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6694b221e7282cc302b7e643c66397b72279398f11ad9550a3602c53ae79c6c
3
+ size 5220
runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5496.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:804a744e01e5818ca8f24738d88b60cfb31ff6c3301152fb790dc50acde63b4b
3
+ size 5220
runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5497.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3c6b1bb64f132d2e83c91d4851516b7b5792c0f0fe446515d23ba80a67056f4
3
+ size 5220
runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5498.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a68921de90fdf17b0001bb4c3c6d96628f31895fbdb5daa18a37f4821c55ddc6
3
+ size 5220
runs/Nov11_20-58-42_gpu113/events.out.tfevents.1762919922.gpu113.5499.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0aa0a54688e10a6a209c75eb0a7879b5b5ea89de9f036e9205ec4672389c02af
3
+ size 5220
runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6519.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c58078e350df532a6b291009e46d4a773d5a0295c204e2b85e06291110780ca8
3
+ size 5220
runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6520.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fb9625cf1062a7a11d63dccb3eeb5a249cf5a62e8ec2becc75d42a5f7f5acef
3
+ size 5220
runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6521.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c40fee9deeca99edbfa9699daae27a3afd271d1dfde293edbd6b2a02671d40b1
3
+ size 5220
runs/Nov11_21-01-33_gpu113/events.out.tfevents.1762920094.gpu113.6522.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e21616cefb0dd7ba54b7baf36ab3b7fe89e15b55efe93fb819351e5899e57cab
3
+ size 5220
runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7427.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9db6fe152b86a449b36ff12977ba390191824289e632996528636a21eaaedbeb
3
+ size 5220
runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7428.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0fe24bc8e8fea326d7befe891556a0ec850125d68de625afae303a313b4ca70
3
+ size 5220
runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7429.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8168f2b866dcf89dd50840054da45731672fde4a0c31203df1e9630283312aab
3
+ size 5220
runs/Nov11_21-03-20_gpu113/events.out.tfevents.1762920200.gpu113.7430.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c838f5237d833596ed3e14ee326f5f338129ea0652c1df125b4ae86c47fe055c
3
+ size 5220
runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8533.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d4d63e1e1577d5dd3e42b01c2777f4ca46dc9e653cc743cb295501e10b1da0a
3
+ size 5220
runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8534.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5664158c3cecf1cbfaa2503e351987dfff3d93ef0fd5c3a37a47d6dc3ca9c27d
3
+ size 5220
runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8535.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aa356e674e6cc7dd9fea4b0cf3de78cdad666f8abb3ffb4bc6f07f11b9dbac9
3
+ size 5220
runs/Nov11_21-07-06_gpu113/events.out.tfevents.1762920427.gpu113.8536.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc81a0546e4a3e7802e64dd8654d5f7b041d225e4a3b9b6f5d5843cb1589049e
3
+ size 5220
runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9634.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bf904334e1c5918bf22fd2193d2b29ca487edfd6aac308444f4ee75bb10ca12
3
+ size 5220
runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9635.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59a7962c858bb2beacf16388deaaa8dac747c540ff520866067e9ab35c07658f
3
+ size 5220
runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9636.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17234bfac74181ca6f5b93ff9ec7210ada0c5f2d439d5f6649ade2a4b7cfa1e9
3
+ size 5220
runs/Nov11_21-10-39_gpu113/events.out.tfevents.1762920639.gpu113.9637.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7532f55ed759273e7d383c331ef59633d8c34609be2e50bb4ed9877a7156597d
3
+ size 5220
runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11623.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fec8be5375539820047bd99c0710f3918bc73a2f38f4787dcc673bac2e249b43
3
+ size 5642
runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11624.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ecb8bc03ea3520cebb264550ae4d9f2d1957f30f150304a767d5bfaad004dfa
3
+ size 5642
runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11625.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1fd60ca38a6ef63ab80c5b94d20c05f8b46f47fe1f90761b25c3fd71e2e0741
3
+ size 5642
runs/Nov11_21-24-27_gpu113/events.out.tfevents.1762921468.gpu113.11626.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb0043fdc1e85821d4777bca0c875e8301fde3b8fff902c287733ad06a9c5f13
3
+ size 5642
runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924909.gpu111.56195.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a130ef8e754c51869d17842104c9a074e02843e929d645cc2beaef9292b2f94d
3
+ size 27016
runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56192.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a92067d8f8ec73a75bf973cc8a42d6c25c072e2728f02194f775d363799b9319
3
+ size 27016
runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56193.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eceab9fba9a467ffe44771758d9ae36c1559a282e007a86191bde764176ab3b1
3
+ size 27016
runs/Nov11_22-21-46_gpu111/events.out.tfevents.1762924910.gpu111.56194.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72df5463e0bc98ec23de918c19ec974ccaa4a090207dd623019656d4e56a0fc7
3
+ size 27016
runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44607.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f504ec0f0ff8e29049188383207dd63b28f25bfc4b35a8c6f06f19a46e4bd56
3
+ size 7118
runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44608.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a85b8974e3f21e0137cac849e2bfd19df7951ca0eb5fb6be12acf861be59998
3
+ size 7118
runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44609.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fa79e6df5424b3051f8578c57ad5c1da199c08cbdd96ae97e514e08e8dbf0b3
3
+ size 7118
runs/Nov12_11-36-10_gpu115/events.out.tfevents.1762972571.gpu115.44610.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c43b8b9327f0587cf2a3391ec3e875d0f6b00806310c770897cba83534e85d94
3
+ size 7118
runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52598.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05a69fcd7b4b083f36df5f453315bc36b858768fce9150e873f4e31e8c00d9eb
3
+ size 7118
runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52599.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06f874f16e4b5ebfcc97d7749cff2cc8d16727ad4cf6f958df08cb720971c8fa
3
+ size 7118
runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52600.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:186202d033d48550787f5e2881707a3a0f7256781f5f88f460e3cf3589214e3b
3
+ size 7118
runs/Nov12_11-36-16_gpu114/events.out.tfevents.1762972577.gpu114.52601.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8130e922ad200617d739abcf23bd3734d9e1a95035bd49965d18267e56d4740c
3
+ size 7329
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a34a66e26624cfd4bf13862a299ad5ec605fb8e78d9b62979723e15d7770596
3
+ size 5368