Hugging Face's logo

timodonnell
/
LlamaFold-experiments

Safetensors
Model card Files
xet
 Community
timodonnell commited on
Commit
0d7f583
·
verified ·
1 Parent(s): dbdbbda

Upload exp5 2B checkpoint (dazzling-universe-12, steps 134k/170k)

Browse files
Files changed (30) hide show
  1. exp5-2b.dazzling-universe-12/checkpoint-134000/config.json +32 -0
  2. exp5-2b.dazzling-universe-12/checkpoint-134000/generation_config.json +10 -0
  3. exp5-2b.dazzling-universe-12/checkpoint-134000/model.safetensors +3 -0
  4. exp5-2b.dazzling-universe-12/checkpoint-134000/optimizer.bin +3 -0
  5. exp5-2b.dazzling-universe-12/checkpoint-134000/pytorch_model_fsdp.bin +3 -0
  6. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_0.pth +3 -0
  7. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_1.pth +3 -0
  8. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_2.pth +3 -0
  9. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_3.pth +3 -0
  10. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_4.pth +3 -0
  11. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_5.pth +3 -0
  12. exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_6.pth +3 -0
  13. exp5-2b.dazzling-universe-12/checkpoint-134000/scheduler.pt +3 -0
  14. exp5-2b.dazzling-universe-12/checkpoint-134000/trainer_state.json +0 -0
  15. exp5-2b.dazzling-universe-12/checkpoint-134000/training_args.bin +3 -0
  16. exp5-2b.dazzling-universe-12/checkpoint-170000/config.json +32 -0
  17. exp5-2b.dazzling-universe-12/checkpoint-170000/generation_config.json +10 -0
  18. exp5-2b.dazzling-universe-12/checkpoint-170000/model.safetensors +3 -0
  19. exp5-2b.dazzling-universe-12/checkpoint-170000/optimizer.bin +3 -0
  20. exp5-2b.dazzling-universe-12/checkpoint-170000/pytorch_model_fsdp.bin +3 -0
  21. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_0.pth +3 -0
  22. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_1.pth +3 -0
  23. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_2.pth +3 -0
  24. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_3.pth +3 -0
  25. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_4.pth +3 -0
  26. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_5.pth +3 -0
  27. exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_6.pth +3 -0
  28. exp5-2b.dazzling-universe-12/checkpoint-170000/scheduler.pt +3 -0
  29. exp5-2b.dazzling-universe-12/checkpoint-170000/trainer_state.json +0 -0
  30. exp5-2b.dazzling-universe-12/checkpoint-170000/training_args.bin +3 -0
exp5-2b.dazzling-universe-12/checkpoint-134000/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "dtype": "float32",
9
+ "eos_token_id": 1,
10
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2048,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 8192,
15
+ "max_position_embeddings": 8192,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 32,
19
+ "num_hidden_layers": 32,
20
+ "num_key_value_heads": 8,
21
+ "pad_token_id": 0,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_parameters": {
25
+ "rope_theta": 10000.0,
26
+ "rope_type": "default"
27
+ },
28
+ "tie_word_embeddings": false,
29
+ "transformers_version": "5.0.0",
30
+ "use_cache": false,
31
+ "vocab_size": 2779
32
+ }
exp5-2b.dazzling-universe-12/checkpoint-134000/generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 1,
5
+ "output_attentions": false,
6
+ "output_hidden_states": false,
7
+ "pad_token_id": 0,
8
+ "transformers_version": "5.0.0",
9
+ "use_cache": true
10
+ }
exp5-2b.dazzling-universe-12/checkpoint-134000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fa06b61c2de9b31d846afcafd34891deabc76ca94fc9cb0d75995c068da641e
3
+ size 7830725176
exp5-2b.dazzling-universe-12/checkpoint-134000/optimizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:951edc0c9317343d110ee58a4e8d93652f0f67d15e905f03b986e681faf2bf94
3
+ size 15661660359
exp5-2b.dazzling-universe-12/checkpoint-134000/pytorch_model_fsdp.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d4894d3c2bd4dc944b8c3b1ef364045989a212e35f61fcb40715a5e0eb546df
3
+ size 7830837376
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa33b409faf0b1aa5dff1988635c613a46afa2a6baba897cea3cad4cec7da83d
3
+ size 16197
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36248ad9443b5c59dc3ab45f7509c5f64e76427877887f6e47621ec6741bc019
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5e80de0e5b4fe95fa8bdaf127ef2c506e5953c5fa7efda1971e53f1dfab4120
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1f8f8fd39e07a78a4f24c3a11a5bfb47d84210e3ef42fcb876f9803fa93d208
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb4d7e8813818cbb91ce749a93d6a2fbd9acf58cf35e1f1eb1ce8c0c0b8d26d3
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:060e20228a62450c5ea7a64962c3545ab4392b1329c0398a9072cdfc42ddb998
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd5f35d14978cc5376bf1af6c30b014c7af192d012079ff9f13e38a0322ed263
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-134000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70c61bea3ee15fbe447492493bb3d6e66d577d23440e0bdce10578561b3a8bec
3
+ size 1465
exp5-2b.dazzling-universe-12/checkpoint-134000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
exp5-2b.dazzling-universe-12/checkpoint-134000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0bc8b491f2cd8a19cbc3fa408e8ec6d2f8cf1c86d3180b0550b0f285003901bc
3
+ size 5649
exp5-2b.dazzling-universe-12/checkpoint-170000/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "dtype": "float32",
9
+ "eos_token_id": 1,
10
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2048,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 8192,
15
+ "max_position_embeddings": 8192,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 32,
19
+ "num_hidden_layers": 32,
20
+ "num_key_value_heads": 8,
21
+ "pad_token_id": 0,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_parameters": {
25
+ "rope_theta": 10000.0,
26
+ "rope_type": "default"
27
+ },
28
+ "tie_word_embeddings": false,
29
+ "transformers_version": "5.0.0",
30
+ "use_cache": false,
31
+ "vocab_size": 2779
32
+ }
exp5-2b.dazzling-universe-12/checkpoint-170000/generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 1,
5
+ "output_attentions": false,
6
+ "output_hidden_states": false,
7
+ "pad_token_id": 0,
8
+ "transformers_version": "5.0.0",
9
+ "use_cache": true
10
+ }
exp5-2b.dazzling-universe-12/checkpoint-170000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3be4430848226dacd5a8fff41a622571312dc52b20634c205f1329538ce62049
3
+ size 7830725176
exp5-2b.dazzling-universe-12/checkpoint-170000/optimizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:803e45b62a425254d6c342c2c07991d88c6958e052855189e8de4dfc11421e2c
3
+ size 15661660359
exp5-2b.dazzling-universe-12/checkpoint-170000/pytorch_model_fsdp.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dfab3977cbd86f773a276d1ed25be2fe8214f7fd6ab954f24098bb0a14912f0
3
+ size 7830837376
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10bd77ca7ba42ccd00779c5d75c7f4b7d7813b88ba44136d7b1f22c25792f832
3
+ size 16197
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e13d3a409536b15f9d490bcce258356c615265fa4971f9b7e59db33335bc8ac4
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:802407269e28cc9b8c104e74ac8d3a0853841ebb55169b49ac1554f4c6bdcaa1
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c8205df426e07855069fa85863856b4955c247708f3b1b35e9c85abb765f3db
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0448a3420605cdb2ecba5d0f7d8c1d9fbc1b45801e62ae88b59d8a64dffb61b7
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76beb9aa1af9cc7aefd63801a3def2009a55196d961829b40911943d3e7b2bb0
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:315c200c386b1ac43ed5f5de4d4dceffee7ac085eb095b7aa522a916cd2f67aa
3
+ size 16133
exp5-2b.dazzling-universe-12/checkpoint-170000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48c53ca5f3f007e586ad4a11007c2fbc820b2a07971c01332b71156d547723e8
3
+ size 1465
exp5-2b.dazzling-universe-12/checkpoint-170000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
exp5-2b.dazzling-universe-12/checkpoint-170000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0bc8b491f2cd8a19cbc3fa408e8ec6d2f8cf1c86d3180b0550b0f285003901bc
3
+ size 5649