Upload folder using huggingface_hub
Browse files- config.json +25 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +51 -0
- tokenizer.json +0 -0
- tokenizer_config.json +71 -0
- trainer_state.json +1243 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 512,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 1024,
|
| 16 |
+
"model_type": "bert",
|
| 17 |
+
"num_attention_heads": 8,
|
| 18 |
+
"num_hidden_layers": 4,
|
| 19 |
+
"pad_token_id": 0,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"transformers_version": "4.57.6",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b9fd14d6ac91d6138ea9fd3ae9ca1d86531d370c51867b90ba95ec353a5e3e5
|
| 3 |
+
size 133031496
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58d9417318d72032075af32d40f8c209d885b2bbbb75ebd7f4abd6b3e46f4ad4
|
| 3 |
+
size 266109515
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c6cf72ff03e07a80741491c7695ae2d4dccd345bd5edfd5ef5e443406b7687b
|
| 3 |
+
size 14645
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88faf51d9670c9f750a337c9c416c4d1fa526f7c096a5b61438fcc6478c89e1a
|
| 3 |
+
size 1465
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "[BOS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "[CLS]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "[EOS]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "[MASK]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "[PAD]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "[SEP]",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "[UNK]",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[UNK]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[CLS]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[SEP]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[PAD]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"30001": {
|
| 44 |
+
"content": "[EOS]",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"30002": {
|
| 52 |
+
"content": "[BOS]",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": false,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": true
|
| 58 |
+
}
|
| 59 |
+
},
|
| 60 |
+
"bos_token": "[BOS]",
|
| 61 |
+
"clean_up_tokenization_spaces": false,
|
| 62 |
+
"cls_token": "[CLS]",
|
| 63 |
+
"eos_token": "[EOS]",
|
| 64 |
+
"extra_special_tokens": {},
|
| 65 |
+
"mask_token": "[MASK]",
|
| 66 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 67 |
+
"pad_token": "[PAD]",
|
| 68 |
+
"sep_token": "[SEP]",
|
| 69 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 70 |
+
"unk_token": "[UNK]"
|
| 71 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1243 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 4000,
|
| 3 |
+
"best_metric": 0.17307986319065094,
|
| 4 |
+
"best_model_checkpoint": "/data/alamparan/mattext_ckpt/results/2026-02-05/01-13-34/pretrain/checkpoints/robocrys_rep_test-pretrain/checkpoint-4000",
|
| 5 |
+
"epoch": 3.875968992248062,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 4000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.04844961240310078,
|
| 14 |
+
"grad_norm": 1.166297197341919,
|
| 15 |
+
"learning_rate": 0.00019981007751937986,
|
| 16 |
+
"loss": 5.9313,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.04844961240310078,
|
| 21 |
+
"eval_loss": 4.05467414855957,
|
| 22 |
+
"eval_runtime": 86.2208,
|
| 23 |
+
"eval_samples_per_second": 220.411,
|
| 24 |
+
"eval_steps_per_second": 4.593,
|
| 25 |
+
"step": 50
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.09689922480620156,
|
| 29 |
+
"grad_norm": 0.7852576375007629,
|
| 30 |
+
"learning_rate": 0.00019961627906976747,
|
| 31 |
+
"loss": 3.8511,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.09689922480620156,
|
| 36 |
+
"eval_loss": 3.614811658859253,
|
| 37 |
+
"eval_runtime": 105.419,
|
| 38 |
+
"eval_samples_per_second": 180.271,
|
| 39 |
+
"eval_steps_per_second": 3.756,
|
| 40 |
+
"step": 100
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.14534883720930233,
|
| 44 |
+
"grad_norm": 0.9356492161750793,
|
| 45 |
+
"learning_rate": 0.00019942248062015505,
|
| 46 |
+
"loss": 3.5665,
|
| 47 |
+
"step": 150
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.14534883720930233,
|
| 51 |
+
"eval_loss": 3.42499041557312,
|
| 52 |
+
"eval_runtime": 119.6728,
|
| 53 |
+
"eval_samples_per_second": 158.8,
|
| 54 |
+
"eval_steps_per_second": 3.309,
|
| 55 |
+
"step": 150
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.1937984496124031,
|
| 59 |
+
"grad_norm": 0.8641079068183899,
|
| 60 |
+
"learning_rate": 0.00019922868217054265,
|
| 61 |
+
"loss": 3.4112,
|
| 62 |
+
"step": 200
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.1937984496124031,
|
| 66 |
+
"eval_loss": 3.2944159507751465,
|
| 67 |
+
"eval_runtime": 107.5464,
|
| 68 |
+
"eval_samples_per_second": 176.705,
|
| 69 |
+
"eval_steps_per_second": 3.682,
|
| 70 |
+
"step": 200
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.24224806201550386,
|
| 74 |
+
"grad_norm": 0.8657866716384888,
|
| 75 |
+
"learning_rate": 0.00019903488372093023,
|
| 76 |
+
"loss": 3.2915,
|
| 77 |
+
"step": 250
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.24224806201550386,
|
| 81 |
+
"eval_loss": 3.1811130046844482,
|
| 82 |
+
"eval_runtime": 111.4126,
|
| 83 |
+
"eval_samples_per_second": 170.573,
|
| 84 |
+
"eval_steps_per_second": 3.554,
|
| 85 |
+
"step": 250
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.29069767441860467,
|
| 89 |
+
"grad_norm": 1.086035132408142,
|
| 90 |
+
"learning_rate": 0.00019884108527131784,
|
| 91 |
+
"loss": 3.2011,
|
| 92 |
+
"step": 300
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.29069767441860467,
|
| 96 |
+
"eval_loss": 3.0944509506225586,
|
| 97 |
+
"eval_runtime": 109.0889,
|
| 98 |
+
"eval_samples_per_second": 174.207,
|
| 99 |
+
"eval_steps_per_second": 3.63,
|
| 100 |
+
"step": 300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.3391472868217054,
|
| 104 |
+
"grad_norm": 1.14714515209198,
|
| 105 |
+
"learning_rate": 0.00019864728682170545,
|
| 106 |
+
"loss": 3.1197,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.3391472868217054,
|
| 111 |
+
"eval_loss": 3.013719081878662,
|
| 112 |
+
"eval_runtime": 113.2854,
|
| 113 |
+
"eval_samples_per_second": 167.753,
|
| 114 |
+
"eval_steps_per_second": 3.496,
|
| 115 |
+
"step": 350
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.3875968992248062,
|
| 119 |
+
"grad_norm": 1.3692086935043335,
|
| 120 |
+
"learning_rate": 0.00019845348837209303,
|
| 121 |
+
"loss": 3.0293,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.3875968992248062,
|
| 126 |
+
"eval_loss": 2.9276790618896484,
|
| 127 |
+
"eval_runtime": 109.3164,
|
| 128 |
+
"eval_samples_per_second": 173.844,
|
| 129 |
+
"eval_steps_per_second": 3.623,
|
| 130 |
+
"step": 400
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.436046511627907,
|
| 134 |
+
"grad_norm": 1.0482667684555054,
|
| 135 |
+
"learning_rate": 0.00019825968992248064,
|
| 136 |
+
"loss": 2.937,
|
| 137 |
+
"step": 450
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.436046511627907,
|
| 141 |
+
"eval_loss": 2.824265718460083,
|
| 142 |
+
"eval_runtime": 113.3929,
|
| 143 |
+
"eval_samples_per_second": 167.594,
|
| 144 |
+
"eval_steps_per_second": 3.492,
|
| 145 |
+
"step": 450
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.4844961240310077,
|
| 149 |
+
"grad_norm": 1.4893537759780884,
|
| 150 |
+
"learning_rate": 0.00019806589147286822,
|
| 151 |
+
"loss": 2.8401,
|
| 152 |
+
"step": 500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.4844961240310077,
|
| 156 |
+
"eval_loss": 2.681960344314575,
|
| 157 |
+
"eval_runtime": 112.2413,
|
| 158 |
+
"eval_samples_per_second": 169.314,
|
| 159 |
+
"eval_steps_per_second": 3.528,
|
| 160 |
+
"step": 500
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.5329457364341085,
|
| 164 |
+
"grad_norm": 1.3668580055236816,
|
| 165 |
+
"learning_rate": 0.00019787209302325582,
|
| 166 |
+
"loss": 2.6953,
|
| 167 |
+
"step": 550
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.5329457364341085,
|
| 171 |
+
"eval_loss": 2.482776165008545,
|
| 172 |
+
"eval_runtime": 109.465,
|
| 173 |
+
"eval_samples_per_second": 173.608,
|
| 174 |
+
"eval_steps_per_second": 3.618,
|
| 175 |
+
"step": 550
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.5813953488372093,
|
| 179 |
+
"grad_norm": 1.7991057634353638,
|
| 180 |
+
"learning_rate": 0.00019767829457364343,
|
| 181 |
+
"loss": 2.4792,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.5813953488372093,
|
| 186 |
+
"eval_loss": 2.0755560398101807,
|
| 187 |
+
"eval_runtime": 111.9949,
|
| 188 |
+
"eval_samples_per_second": 169.686,
|
| 189 |
+
"eval_steps_per_second": 3.536,
|
| 190 |
+
"step": 600
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.6298449612403101,
|
| 194 |
+
"grad_norm": 1.6921905279159546,
|
| 195 |
+
"learning_rate": 0.000197484496124031,
|
| 196 |
+
"loss": 1.9556,
|
| 197 |
+
"step": 650
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.6298449612403101,
|
| 201 |
+
"eval_loss": 1.431490421295166,
|
| 202 |
+
"eval_runtime": 111.5482,
|
| 203 |
+
"eval_samples_per_second": 170.366,
|
| 204 |
+
"eval_steps_per_second": 3.55,
|
| 205 |
+
"step": 650
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.6782945736434108,
|
| 209 |
+
"grad_norm": 1.6035951375961304,
|
| 210 |
+
"learning_rate": 0.00019729069767441862,
|
| 211 |
+
"loss": 1.4683,
|
| 212 |
+
"step": 700
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.6782945736434108,
|
| 216 |
+
"eval_loss": 1.0505075454711914,
|
| 217 |
+
"eval_runtime": 110.2108,
|
| 218 |
+
"eval_samples_per_second": 172.433,
|
| 219 |
+
"eval_steps_per_second": 3.593,
|
| 220 |
+
"step": 700
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.7267441860465116,
|
| 224 |
+
"grad_norm": 1.1903802156448364,
|
| 225 |
+
"learning_rate": 0.0001970968992248062,
|
| 226 |
+
"loss": 1.1292,
|
| 227 |
+
"step": 750
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.7267441860465116,
|
| 231 |
+
"eval_loss": 0.881108820438385,
|
| 232 |
+
"eval_runtime": 112.8095,
|
| 233 |
+
"eval_samples_per_second": 168.461,
|
| 234 |
+
"eval_steps_per_second": 3.51,
|
| 235 |
+
"step": 750
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.7751937984496124,
|
| 239 |
+
"grad_norm": 1.0895042419433594,
|
| 240 |
+
"learning_rate": 0.0001969031007751938,
|
| 241 |
+
"loss": 0.9955,
|
| 242 |
+
"step": 800
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.7751937984496124,
|
| 246 |
+
"eval_loss": 0.7912827730178833,
|
| 247 |
+
"eval_runtime": 113.3802,
|
| 248 |
+
"eval_samples_per_second": 167.613,
|
| 249 |
+
"eval_steps_per_second": 3.493,
|
| 250 |
+
"step": 800
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.8236434108527132,
|
| 254 |
+
"grad_norm": 1.1862001419067383,
|
| 255 |
+
"learning_rate": 0.0001967093023255814,
|
| 256 |
+
"loss": 0.9121,
|
| 257 |
+
"step": 850
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.8236434108527132,
|
| 261 |
+
"eval_loss": 0.7358818650245667,
|
| 262 |
+
"eval_runtime": 110.3147,
|
| 263 |
+
"eval_samples_per_second": 172.271,
|
| 264 |
+
"eval_steps_per_second": 3.59,
|
| 265 |
+
"step": 850
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.872093023255814,
|
| 269 |
+
"grad_norm": 0.9009504318237305,
|
| 270 |
+
"learning_rate": 0.000196515503875969,
|
| 271 |
+
"loss": 0.8213,
|
| 272 |
+
"step": 900
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.872093023255814,
|
| 276 |
+
"eval_loss": 0.673936665058136,
|
| 277 |
+
"eval_runtime": 110.4965,
|
| 278 |
+
"eval_samples_per_second": 171.987,
|
| 279 |
+
"eval_steps_per_second": 3.584,
|
| 280 |
+
"step": 900
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.9205426356589147,
|
| 284 |
+
"grad_norm": 0.8950819373130798,
|
| 285 |
+
"learning_rate": 0.0001963217054263566,
|
| 286 |
+
"loss": 0.7789,
|
| 287 |
+
"step": 950
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.9205426356589147,
|
| 291 |
+
"eval_loss": 0.6290402412414551,
|
| 292 |
+
"eval_runtime": 126.6737,
|
| 293 |
+
"eval_samples_per_second": 150.023,
|
| 294 |
+
"eval_steps_per_second": 3.126,
|
| 295 |
+
"step": 950
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.9689922480620154,
|
| 299 |
+
"grad_norm": 0.8964680433273315,
|
| 300 |
+
"learning_rate": 0.00019612790697674418,
|
| 301 |
+
"loss": 0.7162,
|
| 302 |
+
"step": 1000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.9689922480620154,
|
| 306 |
+
"eval_loss": 0.6034494042396545,
|
| 307 |
+
"eval_runtime": 107.8336,
|
| 308 |
+
"eval_samples_per_second": 176.235,
|
| 309 |
+
"eval_steps_per_second": 3.672,
|
| 310 |
+
"step": 1000
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.0174418604651163,
|
| 314 |
+
"grad_norm": 0.8578282594680786,
|
| 315 |
+
"learning_rate": 0.0001959341085271318,
|
| 316 |
+
"loss": 0.6918,
|
| 317 |
+
"step": 1050
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.0174418604651163,
|
| 321 |
+
"eval_loss": 0.571689784526825,
|
| 322 |
+
"eval_runtime": 109.0331,
|
| 323 |
+
"eval_samples_per_second": 174.296,
|
| 324 |
+
"eval_steps_per_second": 3.632,
|
| 325 |
+
"step": 1050
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.0658914728682172,
|
| 329 |
+
"grad_norm": 0.8138054609298706,
|
| 330 |
+
"learning_rate": 0.00019574031007751937,
|
| 331 |
+
"loss": 0.6908,
|
| 332 |
+
"step": 1100
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.0658914728682172,
|
| 336 |
+
"eval_loss": 0.5466533899307251,
|
| 337 |
+
"eval_runtime": 107.3453,
|
| 338 |
+
"eval_samples_per_second": 177.036,
|
| 339 |
+
"eval_steps_per_second": 3.689,
|
| 340 |
+
"step": 1100
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.1143410852713178,
|
| 344 |
+
"grad_norm": 0.8871294856071472,
|
| 345 |
+
"learning_rate": 0.00019554651162790698,
|
| 346 |
+
"loss": 0.649,
|
| 347 |
+
"step": 1150
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.1143410852713178,
|
| 351 |
+
"eval_loss": 0.523186206817627,
|
| 352 |
+
"eval_runtime": 110.6186,
|
| 353 |
+
"eval_samples_per_second": 171.798,
|
| 354 |
+
"eval_steps_per_second": 3.58,
|
| 355 |
+
"step": 1150
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 1.1627906976744187,
|
| 359 |
+
"grad_norm": 0.7663435339927673,
|
| 360 |
+
"learning_rate": 0.0001953527131782946,
|
| 361 |
+
"loss": 0.5908,
|
| 362 |
+
"step": 1200
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.1627906976744187,
|
| 366 |
+
"eval_loss": 0.5085064172744751,
|
| 367 |
+
"eval_runtime": 108.5509,
|
| 368 |
+
"eval_samples_per_second": 175.07,
|
| 369 |
+
"eval_steps_per_second": 3.648,
|
| 370 |
+
"step": 1200
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 1.2112403100775193,
|
| 374 |
+
"grad_norm": 0.7912157773971558,
|
| 375 |
+
"learning_rate": 0.00019515891472868217,
|
| 376 |
+
"loss": 0.5644,
|
| 377 |
+
"step": 1250
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.2112403100775193,
|
| 381 |
+
"eval_loss": 0.4843023419380188,
|
| 382 |
+
"eval_runtime": 107.8466,
|
| 383 |
+
"eval_samples_per_second": 176.213,
|
| 384 |
+
"eval_steps_per_second": 3.672,
|
| 385 |
+
"step": 1250
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 1.2596899224806202,
|
| 389 |
+
"grad_norm": 0.7282177805900574,
|
| 390 |
+
"learning_rate": 0.00019496511627906978,
|
| 391 |
+
"loss": 0.5577,
|
| 392 |
+
"step": 1300
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.2596899224806202,
|
| 396 |
+
"eval_loss": 0.46449020504951477,
|
| 397 |
+
"eval_runtime": 106.8549,
|
| 398 |
+
"eval_samples_per_second": 177.849,
|
| 399 |
+
"eval_steps_per_second": 3.706,
|
| 400 |
+
"step": 1300
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 1.308139534883721,
|
| 404 |
+
"grad_norm": 0.6605934500694275,
|
| 405 |
+
"learning_rate": 0.00019477131782945736,
|
| 406 |
+
"loss": 0.5335,
|
| 407 |
+
"step": 1350
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.308139534883721,
|
| 411 |
+
"eval_loss": 0.4484124481678009,
|
| 412 |
+
"eval_runtime": 126.0012,
|
| 413 |
+
"eval_samples_per_second": 150.824,
|
| 414 |
+
"eval_steps_per_second": 3.143,
|
| 415 |
+
"step": 1350
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.3565891472868217,
|
| 419 |
+
"grad_norm": 0.6995412707328796,
|
| 420 |
+
"learning_rate": 0.00019457751937984496,
|
| 421 |
+
"loss": 0.5396,
|
| 422 |
+
"step": 1400
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.3565891472868217,
|
| 426 |
+
"eval_loss": 0.4377157688140869,
|
| 427 |
+
"eval_runtime": 118.8732,
|
| 428 |
+
"eval_samples_per_second": 159.868,
|
| 429 |
+
"eval_steps_per_second": 3.331,
|
| 430 |
+
"step": 1400
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.4050387596899225,
|
| 434 |
+
"grad_norm": 0.6287787556648254,
|
| 435 |
+
"learning_rate": 0.00019438372093023257,
|
| 436 |
+
"loss": 0.4988,
|
| 437 |
+
"step": 1450
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.4050387596899225,
|
| 441 |
+
"eval_loss": 0.41537100076675415,
|
| 442 |
+
"eval_runtime": 117.3214,
|
| 443 |
+
"eval_samples_per_second": 161.982,
|
| 444 |
+
"eval_steps_per_second": 3.375,
|
| 445 |
+
"step": 1450
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 1.4534883720930232,
|
| 449 |
+
"grad_norm": 0.7416621446609497,
|
| 450 |
+
"learning_rate": 0.00019418992248062015,
|
| 451 |
+
"loss": 0.4888,
|
| 452 |
+
"step": 1500
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.4534883720930232,
|
| 456 |
+
"eval_loss": 0.4060279130935669,
|
| 457 |
+
"eval_runtime": 118.3486,
|
| 458 |
+
"eval_samples_per_second": 160.577,
|
| 459 |
+
"eval_steps_per_second": 3.346,
|
| 460 |
+
"step": 1500
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.501937984496124,
|
| 464 |
+
"grad_norm": 0.70711749792099,
|
| 465 |
+
"learning_rate": 0.00019399612403100776,
|
| 466 |
+
"loss": 0.4664,
|
| 467 |
+
"step": 1550
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.501937984496124,
|
| 471 |
+
"eval_loss": 0.38495033979415894,
|
| 472 |
+
"eval_runtime": 116.7883,
|
| 473 |
+
"eval_samples_per_second": 162.722,
|
| 474 |
+
"eval_steps_per_second": 3.391,
|
| 475 |
+
"step": 1550
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.550387596899225,
|
| 479 |
+
"grad_norm": 0.6704577207565308,
|
| 480 |
+
"learning_rate": 0.00019380232558139534,
|
| 481 |
+
"loss": 0.4331,
|
| 482 |
+
"step": 1600
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.550387596899225,
|
| 486 |
+
"eval_loss": 0.36785683035850525,
|
| 487 |
+
"eval_runtime": 134.2231,
|
| 488 |
+
"eval_samples_per_second": 141.585,
|
| 489 |
+
"eval_steps_per_second": 2.95,
|
| 490 |
+
"step": 1600
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 1.5988372093023255,
|
| 494 |
+
"grad_norm": 0.7335864305496216,
|
| 495 |
+
"learning_rate": 0.00019360852713178295,
|
| 496 |
+
"loss": 0.4298,
|
| 497 |
+
"step": 1650
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.5988372093023255,
|
| 501 |
+
"eval_loss": 0.34367096424102783,
|
| 502 |
+
"eval_runtime": 193.7622,
|
| 503 |
+
"eval_samples_per_second": 98.079,
|
| 504 |
+
"eval_steps_per_second": 2.044,
|
| 505 |
+
"step": 1650
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.6472868217054264,
|
| 509 |
+
"grad_norm": 0.6677731871604919,
|
| 510 |
+
"learning_rate": 0.00019341472868217055,
|
| 511 |
+
"loss": 0.3978,
|
| 512 |
+
"step": 1700
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.6472868217054264,
|
| 516 |
+
"eval_loss": 0.32630470395088196,
|
| 517 |
+
"eval_runtime": 242.3944,
|
| 518 |
+
"eval_samples_per_second": 78.401,
|
| 519 |
+
"eval_steps_per_second": 1.634,
|
| 520 |
+
"step": 1700
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.695736434108527,
|
| 524 |
+
"grad_norm": 0.8195075392723083,
|
| 525 |
+
"learning_rate": 0.00019322093023255813,
|
| 526 |
+
"loss": 0.3778,
|
| 527 |
+
"step": 1750
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.695736434108527,
|
| 531 |
+
"eval_loss": 0.31104475259780884,
|
| 532 |
+
"eval_runtime": 231.8205,
|
| 533 |
+
"eval_samples_per_second": 81.977,
|
| 534 |
+
"eval_steps_per_second": 1.708,
|
| 535 |
+
"step": 1750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.744186046511628,
|
| 539 |
+
"grad_norm": 0.735170841217041,
|
| 540 |
+
"learning_rate": 0.00019302713178294574,
|
| 541 |
+
"loss": 0.35,
|
| 542 |
+
"step": 1800
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.744186046511628,
|
| 546 |
+
"eval_loss": 0.2956381142139435,
|
| 547 |
+
"eval_runtime": 236.8085,
|
| 548 |
+
"eval_samples_per_second": 80.25,
|
| 549 |
+
"eval_steps_per_second": 1.672,
|
| 550 |
+
"step": 1800
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.7926356589147288,
|
| 554 |
+
"grad_norm": 0.5745160579681396,
|
| 555 |
+
"learning_rate": 0.00019283333333333332,
|
| 556 |
+
"loss": 0.3407,
|
| 557 |
+
"step": 1850
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.7926356589147288,
|
| 561 |
+
"eval_loss": 0.2869073152542114,
|
| 562 |
+
"eval_runtime": 232.6087,
|
| 563 |
+
"eval_samples_per_second": 81.699,
|
| 564 |
+
"eval_steps_per_second": 1.702,
|
| 565 |
+
"step": 1850
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.8410852713178296,
|
| 569 |
+
"grad_norm": 0.6333633661270142,
|
| 570 |
+
"learning_rate": 0.00019263953488372093,
|
| 571 |
+
"loss": 0.3317,
|
| 572 |
+
"step": 1900
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.8410852713178296,
|
| 576 |
+
"eval_loss": 0.2736239731311798,
|
| 577 |
+
"eval_runtime": 235.3635,
|
| 578 |
+
"eval_samples_per_second": 80.743,
|
| 579 |
+
"eval_steps_per_second": 1.683,
|
| 580 |
+
"step": 1900
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.8895348837209303,
|
| 584 |
+
"grad_norm": 0.6406286358833313,
|
| 585 |
+
"learning_rate": 0.00019244573643410854,
|
| 586 |
+
"loss": 0.3247,
|
| 587 |
+
"step": 1950
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.8895348837209303,
|
| 591 |
+
"eval_loss": 0.26805564761161804,
|
| 592 |
+
"eval_runtime": 235.5356,
|
| 593 |
+
"eval_samples_per_second": 80.684,
|
| 594 |
+
"eval_steps_per_second": 1.681,
|
| 595 |
+
"step": 1950
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.937984496124031,
|
| 599 |
+
"grad_norm": 0.4822098910808563,
|
| 600 |
+
"learning_rate": 0.00019225193798449612,
|
| 601 |
+
"loss": 0.3057,
|
| 602 |
+
"step": 2000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.937984496124031,
|
| 606 |
+
"eval_loss": 0.26311546564102173,
|
| 607 |
+
"eval_runtime": 235.1083,
|
| 608 |
+
"eval_samples_per_second": 80.831,
|
| 609 |
+
"eval_steps_per_second": 1.684,
|
| 610 |
+
"step": 2000
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.9864341085271318,
|
| 614 |
+
"grad_norm": 0.6586690545082092,
|
| 615 |
+
"learning_rate": 0.00019205813953488375,
|
| 616 |
+
"loss": 0.2954,
|
| 617 |
+
"step": 2050
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.9864341085271318,
|
| 621 |
+
"eval_loss": 0.2547176778316498,
|
| 622 |
+
"eval_runtime": 235.1319,
|
| 623 |
+
"eval_samples_per_second": 80.823,
|
| 624 |
+
"eval_steps_per_second": 1.684,
|
| 625 |
+
"step": 2050
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 2.0348837209302326,
|
| 629 |
+
"grad_norm": 0.648765504360199,
|
| 630 |
+
"learning_rate": 0.00019186434108527133,
|
| 631 |
+
"loss": 0.3008,
|
| 632 |
+
"step": 2100
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.0348837209302326,
|
| 636 |
+
"eval_loss": 0.2547691762447357,
|
| 637 |
+
"eval_runtime": 227.135,
|
| 638 |
+
"eval_samples_per_second": 83.668,
|
| 639 |
+
"eval_steps_per_second": 1.743,
|
| 640 |
+
"step": 2100
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 2.0833333333333335,
|
| 644 |
+
"grad_norm": 0.6317530870437622,
|
| 645 |
+
"learning_rate": 0.00019167054263565894,
|
| 646 |
+
"loss": 0.2911,
|
| 647 |
+
"step": 2150
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 2.0833333333333335,
|
| 651 |
+
"eval_loss": 0.24894124269485474,
|
| 652 |
+
"eval_runtime": 225.2054,
|
| 653 |
+
"eval_samples_per_second": 84.385,
|
| 654 |
+
"eval_steps_per_second": 1.758,
|
| 655 |
+
"step": 2150
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 2.1317829457364343,
|
| 659 |
+
"grad_norm": 0.5900191068649292,
|
| 660 |
+
"learning_rate": 0.00019147674418604652,
|
| 661 |
+
"loss": 0.2873,
|
| 662 |
+
"step": 2200
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.1317829457364343,
|
| 666 |
+
"eval_loss": 0.24607273936271667,
|
| 667 |
+
"eval_runtime": 231.2435,
|
| 668 |
+
"eval_samples_per_second": 82.182,
|
| 669 |
+
"eval_steps_per_second": 1.712,
|
| 670 |
+
"step": 2200
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 2.1802325581395348,
|
| 674 |
+
"grad_norm": 0.5394904017448425,
|
| 675 |
+
"learning_rate": 0.00019128294573643413,
|
| 676 |
+
"loss": 0.2716,
|
| 677 |
+
"step": 2250
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 2.1802325581395348,
|
| 681 |
+
"eval_loss": 0.2443438619375229,
|
| 682 |
+
"eval_runtime": 225.1673,
|
| 683 |
+
"eval_samples_per_second": 84.399,
|
| 684 |
+
"eval_steps_per_second": 1.759,
|
| 685 |
+
"step": 2250
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 2.2286821705426356,
|
| 689 |
+
"grad_norm": 0.5458412766456604,
|
| 690 |
+
"learning_rate": 0.00019108914728682174,
|
| 691 |
+
"loss": 0.2727,
|
| 692 |
+
"step": 2300
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.2286821705426356,
|
| 696 |
+
"eval_loss": 0.23843063414096832,
|
| 697 |
+
"eval_runtime": 233.0169,
|
| 698 |
+
"eval_samples_per_second": 81.556,
|
| 699 |
+
"eval_steps_per_second": 1.699,
|
| 700 |
+
"step": 2300
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 2.2771317829457365,
|
| 704 |
+
"grad_norm": 0.6243239641189575,
|
| 705 |
+
"learning_rate": 0.00019089534883720932,
|
| 706 |
+
"loss": 0.2741,
|
| 707 |
+
"step": 2350
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 2.2771317829457365,
|
| 711 |
+
"eval_loss": 0.23107607662677765,
|
| 712 |
+
"eval_runtime": 223.8881,
|
| 713 |
+
"eval_samples_per_second": 84.882,
|
| 714 |
+
"eval_steps_per_second": 1.769,
|
| 715 |
+
"step": 2350
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 2.3255813953488373,
|
| 719 |
+
"grad_norm": 0.6094734072685242,
|
| 720 |
+
"learning_rate": 0.00019070155038759692,
|
| 721 |
+
"loss": 0.2719,
|
| 722 |
+
"step": 2400
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.3255813953488373,
|
| 726 |
+
"eval_loss": 0.23176899552345276,
|
| 727 |
+
"eval_runtime": 219.8841,
|
| 728 |
+
"eval_samples_per_second": 86.427,
|
| 729 |
+
"eval_steps_per_second": 1.801,
|
| 730 |
+
"step": 2400
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 2.374031007751938,
|
| 734 |
+
"grad_norm": 0.6149279475212097,
|
| 735 |
+
"learning_rate": 0.0001905077519379845,
|
| 736 |
+
"loss": 0.2641,
|
| 737 |
+
"step": 2450
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 2.374031007751938,
|
| 741 |
+
"eval_loss": 0.22322338819503784,
|
| 742 |
+
"eval_runtime": 224.8759,
|
| 743 |
+
"eval_samples_per_second": 84.509,
|
| 744 |
+
"eval_steps_per_second": 1.761,
|
| 745 |
+
"step": 2450
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.4224806201550386,
|
| 749 |
+
"grad_norm": 0.5470075607299805,
|
| 750 |
+
"learning_rate": 0.0001903139534883721,
|
| 751 |
+
"loss": 0.2558,
|
| 752 |
+
"step": 2500
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.4224806201550386,
|
| 756 |
+
"eval_loss": 0.22380074858665466,
|
| 757 |
+
"eval_runtime": 220.6655,
|
| 758 |
+
"eval_samples_per_second": 86.121,
|
| 759 |
+
"eval_steps_per_second": 1.795,
|
| 760 |
+
"step": 2500
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 2.4709302325581395,
|
| 764 |
+
"grad_norm": 0.6210835576057434,
|
| 765 |
+
"learning_rate": 0.00019012015503875972,
|
| 766 |
+
"loss": 0.2544,
|
| 767 |
+
"step": 2550
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.4709302325581395,
|
| 771 |
+
"eval_loss": 0.22070536017417908,
|
| 772 |
+
"eval_runtime": 223.1476,
|
| 773 |
+
"eval_samples_per_second": 85.163,
|
| 774 |
+
"eval_steps_per_second": 1.775,
|
| 775 |
+
"step": 2550
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 2.5193798449612403,
|
| 779 |
+
"grad_norm": 0.43341922760009766,
|
| 780 |
+
"learning_rate": 0.0001899263565891473,
|
| 781 |
+
"loss": 0.2496,
|
| 782 |
+
"step": 2600
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.5193798449612403,
|
| 786 |
+
"eval_loss": 0.2163931280374527,
|
| 787 |
+
"eval_runtime": 232.0825,
|
| 788 |
+
"eval_samples_per_second": 81.885,
|
| 789 |
+
"eval_steps_per_second": 1.706,
|
| 790 |
+
"step": 2600
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 2.567829457364341,
|
| 794 |
+
"grad_norm": 0.4871957302093506,
|
| 795 |
+
"learning_rate": 0.0001897325581395349,
|
| 796 |
+
"loss": 0.2506,
|
| 797 |
+
"step": 2650
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 2.567829457364341,
|
| 801 |
+
"eval_loss": 0.21461744606494904,
|
| 802 |
+
"eval_runtime": 219.2336,
|
| 803 |
+
"eval_samples_per_second": 86.684,
|
| 804 |
+
"eval_steps_per_second": 1.806,
|
| 805 |
+
"step": 2650
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 2.616279069767442,
|
| 809 |
+
"grad_norm": 0.5310043692588806,
|
| 810 |
+
"learning_rate": 0.0001895387596899225,
|
| 811 |
+
"loss": 0.247,
|
| 812 |
+
"step": 2700
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.616279069767442,
|
| 816 |
+
"eval_loss": 0.21458372473716736,
|
| 817 |
+
"eval_runtime": 224.6247,
|
| 818 |
+
"eval_samples_per_second": 84.603,
|
| 819 |
+
"eval_steps_per_second": 1.763,
|
| 820 |
+
"step": 2700
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 2.6647286821705425,
|
| 824 |
+
"grad_norm": 0.6074191927909851,
|
| 825 |
+
"learning_rate": 0.0001893449612403101,
|
| 826 |
+
"loss": 0.2455,
|
| 827 |
+
"step": 2750
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 2.6647286821705425,
|
| 831 |
+
"eval_loss": 0.21306496858596802,
|
| 832 |
+
"eval_runtime": 223.479,
|
| 833 |
+
"eval_samples_per_second": 85.037,
|
| 834 |
+
"eval_steps_per_second": 1.772,
|
| 835 |
+
"step": 2750
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 2.7131782945736433,
|
| 839 |
+
"grad_norm": 0.5458905100822449,
|
| 840 |
+
"learning_rate": 0.00018915116279069768,
|
| 841 |
+
"loss": 0.243,
|
| 842 |
+
"step": 2800
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.7131782945736433,
|
| 846 |
+
"eval_loss": 0.2097616195678711,
|
| 847 |
+
"eval_runtime": 220.5231,
|
| 848 |
+
"eval_samples_per_second": 86.177,
|
| 849 |
+
"eval_steps_per_second": 1.796,
|
| 850 |
+
"step": 2800
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 2.761627906976744,
|
| 854 |
+
"grad_norm": 0.4153260588645935,
|
| 855 |
+
"learning_rate": 0.00018895736434108528,
|
| 856 |
+
"loss": 0.2356,
|
| 857 |
+
"step": 2850
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.761627906976744,
|
| 861 |
+
"eval_loss": 0.20619072020053864,
|
| 862 |
+
"eval_runtime": 214.0517,
|
| 863 |
+
"eval_samples_per_second": 88.782,
|
| 864 |
+
"eval_steps_per_second": 1.85,
|
| 865 |
+
"step": 2850
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 2.810077519379845,
|
| 869 |
+
"grad_norm": 0.4979017674922943,
|
| 870 |
+
"learning_rate": 0.0001887635658914729,
|
| 871 |
+
"loss": 0.2361,
|
| 872 |
+
"step": 2900
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.810077519379845,
|
| 876 |
+
"eval_loss": 0.20226989686489105,
|
| 877 |
+
"eval_runtime": 216.8896,
|
| 878 |
+
"eval_samples_per_second": 87.621,
|
| 879 |
+
"eval_steps_per_second": 1.826,
|
| 880 |
+
"step": 2900
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 2.858527131782946,
|
| 884 |
+
"grad_norm": 0.5476083755493164,
|
| 885 |
+
"learning_rate": 0.00018856976744186047,
|
| 886 |
+
"loss": 0.2347,
|
| 887 |
+
"step": 2950
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 2.858527131782946,
|
| 891 |
+
"eval_loss": 0.2028292417526245,
|
| 892 |
+
"eval_runtime": 214.8481,
|
| 893 |
+
"eval_samples_per_second": 88.453,
|
| 894 |
+
"eval_steps_per_second": 1.843,
|
| 895 |
+
"step": 2950
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 2.9069767441860463,
|
| 899 |
+
"grad_norm": 0.5912747979164124,
|
| 900 |
+
"learning_rate": 0.00018837596899224808,
|
| 901 |
+
"loss": 0.2367,
|
| 902 |
+
"step": 3000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 2.9069767441860463,
|
| 906 |
+
"eval_loss": 0.1990610510110855,
|
| 907 |
+
"eval_runtime": 219.8884,
|
| 908 |
+
"eval_samples_per_second": 86.426,
|
| 909 |
+
"eval_steps_per_second": 1.801,
|
| 910 |
+
"step": 3000
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 2.955426356589147,
|
| 914 |
+
"grad_norm": 0.4696279466152191,
|
| 915 |
+
"learning_rate": 0.00018818217054263566,
|
| 916 |
+
"loss": 0.2292,
|
| 917 |
+
"step": 3050
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 2.955426356589147,
|
| 921 |
+
"eval_loss": 0.20148637890815735,
|
| 922 |
+
"eval_runtime": 225.0994,
|
| 923 |
+
"eval_samples_per_second": 84.425,
|
| 924 |
+
"eval_steps_per_second": 1.759,
|
| 925 |
+
"step": 3050
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 3.003875968992248,
|
| 929 |
+
"grad_norm": 0.47198325395584106,
|
| 930 |
+
"learning_rate": 0.00018798837209302327,
|
| 931 |
+
"loss": 0.2286,
|
| 932 |
+
"step": 3100
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 3.003875968992248,
|
| 936 |
+
"eval_loss": 0.1965794712305069,
|
| 937 |
+
"eval_runtime": 216.7977,
|
| 938 |
+
"eval_samples_per_second": 87.658,
|
| 939 |
+
"eval_steps_per_second": 1.827,
|
| 940 |
+
"step": 3100
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 3.052325581395349,
|
| 944 |
+
"grad_norm": 0.49365678429603577,
|
| 945 |
+
"learning_rate": 0.00018779457364341087,
|
| 946 |
+
"loss": 0.2239,
|
| 947 |
+
"step": 3150
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 3.052325581395349,
|
| 951 |
+
"eval_loss": 0.1976451575756073,
|
| 952 |
+
"eval_runtime": 215.2246,
|
| 953 |
+
"eval_samples_per_second": 88.298,
|
| 954 |
+
"eval_steps_per_second": 1.84,
|
| 955 |
+
"step": 3150
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 3.10077519379845,
|
| 959 |
+
"grad_norm": 0.5356510281562805,
|
| 960 |
+
"learning_rate": 0.00018760077519379845,
|
| 961 |
+
"loss": 0.224,
|
| 962 |
+
"step": 3200
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 3.10077519379845,
|
| 966 |
+
"eval_loss": 0.1925686150789261,
|
| 967 |
+
"eval_runtime": 218.8392,
|
| 968 |
+
"eval_samples_per_second": 86.84,
|
| 969 |
+
"eval_steps_per_second": 1.81,
|
| 970 |
+
"step": 3200
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 3.14922480620155,
|
| 974 |
+
"grad_norm": 0.46862003207206726,
|
| 975 |
+
"learning_rate": 0.00018740697674418606,
|
| 976 |
+
"loss": 0.212,
|
| 977 |
+
"step": 3250
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 3.14922480620155,
|
| 981 |
+
"eval_loss": 0.19242693483829498,
|
| 982 |
+
"eval_runtime": 216.1261,
|
| 983 |
+
"eval_samples_per_second": 87.93,
|
| 984 |
+
"eval_steps_per_second": 1.832,
|
| 985 |
+
"step": 3250
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 3.197674418604651,
|
| 989 |
+
"grad_norm": 0.40870046615600586,
|
| 990 |
+
"learning_rate": 0.00018721317829457364,
|
| 991 |
+
"loss": 0.2188,
|
| 992 |
+
"step": 3300
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 3.197674418604651,
|
| 996 |
+
"eval_loss": 0.18868175148963928,
|
| 997 |
+
"eval_runtime": 214.9811,
|
| 998 |
+
"eval_samples_per_second": 88.398,
|
| 999 |
+
"eval_steps_per_second": 1.842,
|
| 1000 |
+
"step": 3300
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 3.246124031007752,
|
| 1004 |
+
"grad_norm": 0.5424318909645081,
|
| 1005 |
+
"learning_rate": 0.00018701937984496125,
|
| 1006 |
+
"loss": 0.2157,
|
| 1007 |
+
"step": 3350
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 3.246124031007752,
|
| 1011 |
+
"eval_loss": 0.1919330656528473,
|
| 1012 |
+
"eval_runtime": 217.7522,
|
| 1013 |
+
"eval_samples_per_second": 87.274,
|
| 1014 |
+
"eval_steps_per_second": 1.819,
|
| 1015 |
+
"step": 3350
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 3.294573643410853,
|
| 1019 |
+
"grad_norm": 0.41225236654281616,
|
| 1020 |
+
"learning_rate": 0.00018682558139534886,
|
| 1021 |
+
"loss": 0.2156,
|
| 1022 |
+
"step": 3400
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 3.294573643410853,
|
| 1026 |
+
"eval_loss": 0.18903516232967377,
|
| 1027 |
+
"eval_runtime": 221.615,
|
| 1028 |
+
"eval_samples_per_second": 85.752,
|
| 1029 |
+
"eval_steps_per_second": 1.787,
|
| 1030 |
+
"step": 3400
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 3.3430232558139537,
|
| 1034 |
+
"grad_norm": 0.4955701231956482,
|
| 1035 |
+
"learning_rate": 0.00018663178294573644,
|
| 1036 |
+
"loss": 0.2087,
|
| 1037 |
+
"step": 3450
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 3.3430232558139537,
|
| 1041 |
+
"eval_loss": 0.18453241884708405,
|
| 1042 |
+
"eval_runtime": 217.9492,
|
| 1043 |
+
"eval_samples_per_second": 87.195,
|
| 1044 |
+
"eval_steps_per_second": 1.817,
|
| 1045 |
+
"step": 3450
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 3.391472868217054,
|
| 1049 |
+
"grad_norm": 0.4107048809528351,
|
| 1050 |
+
"learning_rate": 0.00018643798449612405,
|
| 1051 |
+
"loss": 0.2128,
|
| 1052 |
+
"step": 3500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 3.391472868217054,
|
| 1056 |
+
"eval_loss": 0.18335753679275513,
|
| 1057 |
+
"eval_runtime": 222.918,
|
| 1058 |
+
"eval_samples_per_second": 85.251,
|
| 1059 |
+
"eval_steps_per_second": 1.776,
|
| 1060 |
+
"step": 3500
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 3.439922480620155,
|
| 1064 |
+
"grad_norm": 0.44138869643211365,
|
| 1065 |
+
"learning_rate": 0.00018624418604651163,
|
| 1066 |
+
"loss": 0.2069,
|
| 1067 |
+
"step": 3550
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 3.439922480620155,
|
| 1071 |
+
"eval_loss": 0.18529056012630463,
|
| 1072 |
+
"eval_runtime": 226.8247,
|
| 1073 |
+
"eval_samples_per_second": 83.783,
|
| 1074 |
+
"eval_steps_per_second": 1.746,
|
| 1075 |
+
"step": 3550
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 3.488372093023256,
|
| 1079 |
+
"grad_norm": 0.41599947214126587,
|
| 1080 |
+
"learning_rate": 0.00018605038759689923,
|
| 1081 |
+
"loss": 0.2069,
|
| 1082 |
+
"step": 3600
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 3.488372093023256,
|
| 1086 |
+
"eval_loss": 0.18560050427913666,
|
| 1087 |
+
"eval_runtime": 237.1428,
|
| 1088 |
+
"eval_samples_per_second": 80.137,
|
| 1089 |
+
"eval_steps_per_second": 1.67,
|
| 1090 |
+
"step": 3600
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 3.5368217054263567,
|
| 1094 |
+
"grad_norm": 0.39731481671333313,
|
| 1095 |
+
"learning_rate": 0.00018585658914728684,
|
| 1096 |
+
"loss": 0.2074,
|
| 1097 |
+
"step": 3650
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 3.5368217054263567,
|
| 1101 |
+
"eval_loss": 0.1789853423833847,
|
| 1102 |
+
"eval_runtime": 220.4447,
|
| 1103 |
+
"eval_samples_per_second": 86.208,
|
| 1104 |
+
"eval_steps_per_second": 1.796,
|
| 1105 |
+
"step": 3650
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 3.5852713178294575,
|
| 1109 |
+
"grad_norm": 0.46680784225463867,
|
| 1110 |
+
"learning_rate": 0.00018566279069767442,
|
| 1111 |
+
"loss": 0.2042,
|
| 1112 |
+
"step": 3700
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 3.5852713178294575,
|
| 1116 |
+
"eval_loss": 0.1809505820274353,
|
| 1117 |
+
"eval_runtime": 218.2028,
|
| 1118 |
+
"eval_samples_per_second": 87.093,
|
| 1119 |
+
"eval_steps_per_second": 1.815,
|
| 1120 |
+
"step": 3700
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 3.633720930232558,
|
| 1124 |
+
"grad_norm": 0.552648663520813,
|
| 1125 |
+
"learning_rate": 0.00018546899224806203,
|
| 1126 |
+
"loss": 0.2089,
|
| 1127 |
+
"step": 3750
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 3.633720930232558,
|
| 1131 |
+
"eval_loss": 0.17837607860565186,
|
| 1132 |
+
"eval_runtime": 221.244,
|
| 1133 |
+
"eval_samples_per_second": 85.896,
|
| 1134 |
+
"eval_steps_per_second": 1.79,
|
| 1135 |
+
"step": 3750
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 3.682170542635659,
|
| 1139 |
+
"grad_norm": 0.5326802134513855,
|
| 1140 |
+
"learning_rate": 0.0001852751937984496,
|
| 1141 |
+
"loss": 0.1985,
|
| 1142 |
+
"step": 3800
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 3.682170542635659,
|
| 1146 |
+
"eval_loss": 0.17835576832294464,
|
| 1147 |
+
"eval_runtime": 227.413,
|
| 1148 |
+
"eval_samples_per_second": 83.566,
|
| 1149 |
+
"eval_steps_per_second": 1.741,
|
| 1150 |
+
"step": 3800
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 3.7306201550387597,
|
| 1154 |
+
"grad_norm": 0.5640744566917419,
|
| 1155 |
+
"learning_rate": 0.00018508139534883722,
|
| 1156 |
+
"loss": 0.201,
|
| 1157 |
+
"step": 3850
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 3.7306201550387597,
|
| 1161 |
+
"eval_loss": 0.17722435295581818,
|
| 1162 |
+
"eval_runtime": 220.9949,
|
| 1163 |
+
"eval_samples_per_second": 85.993,
|
| 1164 |
+
"eval_steps_per_second": 1.792,
|
| 1165 |
+
"step": 3850
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 3.7790697674418605,
|
| 1169 |
+
"grad_norm": 0.41974276304244995,
|
| 1170 |
+
"learning_rate": 0.00018488759689922482,
|
| 1171 |
+
"loss": 0.1945,
|
| 1172 |
+
"step": 3900
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 3.7790697674418605,
|
| 1176 |
+
"eval_loss": 0.17568857967853546,
|
| 1177 |
+
"eval_runtime": 227.7226,
|
| 1178 |
+
"eval_samples_per_second": 83.452,
|
| 1179 |
+
"eval_steps_per_second": 1.739,
|
| 1180 |
+
"step": 3900
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 3.8275193798449614,
|
| 1184 |
+
"grad_norm": 0.4548977315425873,
|
| 1185 |
+
"learning_rate": 0.0001846937984496124,
|
| 1186 |
+
"loss": 0.2035,
|
| 1187 |
+
"step": 3950
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 3.8275193798449614,
|
| 1191 |
+
"eval_loss": 0.17479188740253448,
|
| 1192 |
+
"eval_runtime": 225.9318,
|
| 1193 |
+
"eval_samples_per_second": 84.114,
|
| 1194 |
+
"eval_steps_per_second": 1.753,
|
| 1195 |
+
"step": 3950
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 3.875968992248062,
|
| 1199 |
+
"grad_norm": 0.44886454939842224,
|
| 1200 |
+
"learning_rate": 0.0001845,
|
| 1201 |
+
"loss": 0.1965,
|
| 1202 |
+
"step": 4000
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 3.875968992248062,
|
| 1206 |
+
"eval_loss": 0.17307986319065094,
|
| 1207 |
+
"eval_runtime": 221.697,
|
| 1208 |
+
"eval_samples_per_second": 85.721,
|
| 1209 |
+
"eval_steps_per_second": 1.786,
|
| 1210 |
+
"step": 4000
|
| 1211 |
+
}
|
| 1212 |
+
],
|
| 1213 |
+
"logging_steps": 50,
|
| 1214 |
+
"max_steps": 51600,
|
| 1215 |
+
"num_input_tokens_seen": 0,
|
| 1216 |
+
"num_train_epochs": 50,
|
| 1217 |
+
"save_steps": 1000,
|
| 1218 |
+
"stateful_callbacks": {
|
| 1219 |
+
"EarlyStoppingCallback": {
|
| 1220 |
+
"args": {
|
| 1221 |
+
"early_stopping_patience": 10,
|
| 1222 |
+
"early_stopping_threshold": 0.001
|
| 1223 |
+
},
|
| 1224 |
+
"attributes": {
|
| 1225 |
+
"early_stopping_patience_counter": 0
|
| 1226 |
+
}
|
| 1227 |
+
},
|
| 1228 |
+
"TrainerControl": {
|
| 1229 |
+
"args": {
|
| 1230 |
+
"should_epoch_stop": false,
|
| 1231 |
+
"should_evaluate": false,
|
| 1232 |
+
"should_log": false,
|
| 1233 |
+
"should_save": true,
|
| 1234 |
+
"should_training_stop": false
|
| 1235 |
+
},
|
| 1236 |
+
"attributes": {}
|
| 1237 |
+
}
|
| 1238 |
+
},
|
| 1239 |
+
"total_flos": 4.034795399365018e+16,
|
| 1240 |
+
"train_batch_size": 96,
|
| 1241 |
+
"trial_name": null,
|
| 1242 |
+
"trial_params": null
|
| 1243 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae525e96f7b11a1fae7d5898eb3b17684471e4365a62484304102b6f4c3c18fe
|
| 3 |
+
size 15761
|