Upload fine-tuned CodeModernBERT-GPT2 model for code-to-text
Browse files- config.json +100 -0
- generation_config.json +5 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +6 -0
- tokenizer.json +0 -0
- tokenizer_config.json +21 -0
- trainer_state.json +756 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"EncoderDecoderModel"
|
| 4 |
+
],
|
| 5 |
+
"cache_dir": "./cache",
|
| 6 |
+
"decoder": {
|
| 7 |
+
"_attn_implementation_autoset": true,
|
| 8 |
+
"_name_or_path": "openai-community/gpt2-large",
|
| 9 |
+
"activation_function": "gelu_new",
|
| 10 |
+
"add_cross_attention": true,
|
| 11 |
+
"architectures": [
|
| 12 |
+
"GPT2LMHeadModel"
|
| 13 |
+
],
|
| 14 |
+
"attn_pdrop": 0.1,
|
| 15 |
+
"embd_pdrop": 0.1,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"is_decoder": true,
|
| 18 |
+
"layer_norm_epsilon": 1e-05,
|
| 19 |
+
"model_type": "gpt2",
|
| 20 |
+
"n_ctx": 1024,
|
| 21 |
+
"n_embd": 1280,
|
| 22 |
+
"n_head": 20,
|
| 23 |
+
"n_inner": null,
|
| 24 |
+
"n_layer": 36,
|
| 25 |
+
"n_positions": 1024,
|
| 26 |
+
"reorder_and_upcast_attn": false,
|
| 27 |
+
"resid_pdrop": 0.1,
|
| 28 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 29 |
+
"scale_attn_weights": true,
|
| 30 |
+
"summary_activation": null,
|
| 31 |
+
"summary_first_dropout": 0.1,
|
| 32 |
+
"summary_proj_to_labels": true,
|
| 33 |
+
"summary_type": "cls_index",
|
| 34 |
+
"summary_use_proj": true,
|
| 35 |
+
"task_specific_params": {
|
| 36 |
+
"text-generation": {
|
| 37 |
+
"do_sample": true,
|
| 38 |
+
"max_length": 50
|
| 39 |
+
}
|
| 40 |
+
},
|
| 41 |
+
"torch_dtype": "float32",
|
| 42 |
+
"use_cache": true,
|
| 43 |
+
"vocab_size": 50257
|
| 44 |
+
},
|
| 45 |
+
"decoder_start_token_id": 50256,
|
| 46 |
+
"encoder": {
|
| 47 |
+
"_attn_implementation_autoset": true,
|
| 48 |
+
"_name_or_path": "Shuu12121/CodeModernBERT-Ghost",
|
| 49 |
+
"architectures": [
|
| 50 |
+
"ModernBertForMaskedLM"
|
| 51 |
+
],
|
| 52 |
+
"attention_bias": false,
|
| 53 |
+
"attention_dropout": 0.0,
|
| 54 |
+
"attention_probs_dropout_prob": 0.1,
|
| 55 |
+
"bos_token_id": 50000,
|
| 56 |
+
"classifier_activation": "gelu",
|
| 57 |
+
"classifier_bias": false,
|
| 58 |
+
"classifier_dropout": 0.0,
|
| 59 |
+
"classifier_pooling": "cls",
|
| 60 |
+
"cls_token_id": 50281,
|
| 61 |
+
"decoder_bias": true,
|
| 62 |
+
"deterministic_flash_attn": false,
|
| 63 |
+
"embedding_dropout": 0.0,
|
| 64 |
+
"eos_token_id": 50001,
|
| 65 |
+
"global_attn_every_n_layers": 3,
|
| 66 |
+
"global_rope_theta": 160000.0,
|
| 67 |
+
"hidden_activation": "gelu",
|
| 68 |
+
"hidden_dropout_prob": 0.1,
|
| 69 |
+
"hidden_size": 768,
|
| 70 |
+
"initializer_cutoff_factor": 2.0,
|
| 71 |
+
"initializer_range": 0.02,
|
| 72 |
+
"intermediate_size": 3072,
|
| 73 |
+
"local_attention": 128,
|
| 74 |
+
"local_attention_rope_theta": 10000,
|
| 75 |
+
"local_attention_window": 128,
|
| 76 |
+
"local_rope_theta": 10000.0,
|
| 77 |
+
"max_position_embeddings": 2048,
|
| 78 |
+
"mlp_bias": false,
|
| 79 |
+
"mlp_dropout": 0.0,
|
| 80 |
+
"model_type": "modernbert",
|
| 81 |
+
"norm_bias": false,
|
| 82 |
+
"norm_eps": 1e-05,
|
| 83 |
+
"num_attention_heads": 12,
|
| 84 |
+
"num_hidden_layers": 12,
|
| 85 |
+
"pad_token_id": 0,
|
| 86 |
+
"repad_logits_with_grad": false,
|
| 87 |
+
"rope_theta": 160000,
|
| 88 |
+
"sparse_pred_ignore_index": -100,
|
| 89 |
+
"sparse_prediction": false,
|
| 90 |
+
"torch_dtype": "float32",
|
| 91 |
+
"type_vocab_size": 2,
|
| 92 |
+
"vocab_size": 50004
|
| 93 |
+
},
|
| 94 |
+
"eos_token_id": 50256,
|
| 95 |
+
"is_encoder_decoder": true,
|
| 96 |
+
"model_type": "encoder-decoder",
|
| 97 |
+
"pad_token_id": 50256,
|
| 98 |
+
"torch_dtype": "float32",
|
| 99 |
+
"transformers_version": "4.51.1"
|
| 100 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 50256,
|
| 3 |
+
"eos_token_id": 50256,
|
| 4 |
+
"transformers_version": "4.51.1"
|
| 5 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86eee688750922fac48efea5463f0030cee4d85dc47c24290a33e6fa369a20d2
|
| 3 |
+
size 4651648824
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2df0f1fe54f49d765beaa6da4a951e92a0b708b96a3c1ff497c34cec39293df1
|
| 3 |
+
size 9303790734
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fd4fbd0a0af51fce0346337b86feeb70097c8160d5f799a45640c69ced4dfe6
|
| 3 |
+
size 14244
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8e1b543379bebe1f4040c969f528e4856e990a9c6ed611dd98a2ffc02ed4612
|
| 3 |
+
size 988
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4040b3d3a37d5268a49213c4aab6920b75793eec2594edd73004e146f6f3a025
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<|endoftext|>",
|
| 3 |
+
"eos_token": "<|endoftext|>",
|
| 4 |
+
"pad_token": "<|endoftext|>",
|
| 5 |
+
"unk_token": "<|endoftext|>"
|
| 6 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"50256": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
}
|
| 12 |
+
},
|
| 13 |
+
"bos_token": "<|endoftext|>",
|
| 14 |
+
"clean_up_tokenization_spaces": false,
|
| 15 |
+
"eos_token": "<|endoftext|>",
|
| 16 |
+
"extra_special_tokens": {},
|
| 17 |
+
"model_max_length": 1024,
|
| 18 |
+
"pad_token": "<|endoftext|>",
|
| 19 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 20 |
+
"unk_token": "<|endoftext|>"
|
| 21 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,756 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 10000,
|
| 3 |
+
"best_metric": 0.0,
|
| 4 |
+
"best_model_checkpoint": "./results_code_to_text_codemodernbert_gpt2_all_languages/checkpoint-10000",
|
| 5 |
+
"epoch": 1.151688790348124,
|
| 6 |
+
"eval_steps": 10000,
|
| 7 |
+
"global_step": 10000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.011517751734861355,
|
| 14 |
+
"grad_norm": 2.149099826812744,
|
| 15 |
+
"learning_rate": 9.999664806726203e-05,
|
| 16 |
+
"loss": 7.9953,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.02303550346972271,
|
| 21 |
+
"grad_norm": 2.219588041305542,
|
| 22 |
+
"learning_rate": 9.998602828889252e-05,
|
| 23 |
+
"loss": 4.0086,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03455325520458406,
|
| 28 |
+
"grad_norm": 2.8714349269866943,
|
| 29 |
+
"learning_rate": 9.996813639049668e-05,
|
| 30 |
+
"loss": 3.5128,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.04607100693944542,
|
| 35 |
+
"grad_norm": 2.233255386352539,
|
| 36 |
+
"learning_rate": 9.994297497504253e-05,
|
| 37 |
+
"loss": 2.5397,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.057588758674306774,
|
| 42 |
+
"grad_norm": 3.247396469116211,
|
| 43 |
+
"learning_rate": 9.991054770308961e-05,
|
| 44 |
+
"loss": 1.9996,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.06910651040916813,
|
| 49 |
+
"grad_norm": 3.2835168838500977,
|
| 50 |
+
"learning_rate": 9.987085929225645e-05,
|
| 51 |
+
"loss": 1.7329,
|
| 52 |
+
"step": 600
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.08062426214402948,
|
| 56 |
+
"grad_norm": 2.6005859375,
|
| 57 |
+
"learning_rate": 9.98239155165343e-05,
|
| 58 |
+
"loss": 1.6232,
|
| 59 |
+
"step": 700
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.09214201387889084,
|
| 63 |
+
"grad_norm": 1.9255400896072388,
|
| 64 |
+
"learning_rate": 9.976972320544693e-05,
|
| 65 |
+
"loss": 1.5533,
|
| 66 |
+
"step": 800
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.10365976561375219,
|
| 70 |
+
"grad_norm": 2.3379406929016113,
|
| 71 |
+
"learning_rate": 9.970829024305725e-05,
|
| 72 |
+
"loss": 1.5081,
|
| 73 |
+
"step": 900
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.11517751734861355,
|
| 77 |
+
"grad_norm": 1.4015233516693115,
|
| 78 |
+
"learning_rate": 9.963962556682021e-05,
|
| 79 |
+
"loss": 1.4448,
|
| 80 |
+
"step": 1000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.1266952690834749,
|
| 84 |
+
"grad_norm": 2.204472541809082,
|
| 85 |
+
"learning_rate": 9.956373916628259e-05,
|
| 86 |
+
"loss": 1.3501,
|
| 87 |
+
"step": 1100
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.13821302081833625,
|
| 91 |
+
"grad_norm": 2.014930248260498,
|
| 92 |
+
"learning_rate": 9.948064208162966e-05,
|
| 93 |
+
"loss": 1.3379,
|
| 94 |
+
"step": 1200
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.1497307725531976,
|
| 98 |
+
"grad_norm": 1.2242907285690308,
|
| 99 |
+
"learning_rate": 9.939034640207901e-05,
|
| 100 |
+
"loss": 1.3679,
|
| 101 |
+
"step": 1300
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.16124852428805897,
|
| 105 |
+
"grad_norm": 1.4741770029067993,
|
| 106 |
+
"learning_rate": 9.929286526412182e-05,
|
| 107 |
+
"loss": 1.2859,
|
| 108 |
+
"step": 1400
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.17276627602292033,
|
| 112 |
+
"grad_norm": 1.6436251401901245,
|
| 113 |
+
"learning_rate": 9.91882128496117e-05,
|
| 114 |
+
"loss": 1.2964,
|
| 115 |
+
"step": 1500
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.1842840277577817,
|
| 119 |
+
"grad_norm": 1.5630017518997192,
|
| 120 |
+
"learning_rate": 9.907640438370148e-05,
|
| 121 |
+
"loss": 1.3205,
|
| 122 |
+
"step": 1600
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.19580177949264305,
|
| 126 |
+
"grad_norm": 1.3735089302062988,
|
| 127 |
+
"learning_rate": 9.895745613262816e-05,
|
| 128 |
+
"loss": 1.2516,
|
| 129 |
+
"step": 1700
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.20731953122750438,
|
| 133 |
+
"grad_norm": 1.54973566532135,
|
| 134 |
+
"learning_rate": 9.883138540134653e-05,
|
| 135 |
+
"loss": 1.2208,
|
| 136 |
+
"step": 1800
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.21883728296236574,
|
| 140 |
+
"grad_norm": 1.3270553350448608,
|
| 141 |
+
"learning_rate": 9.869821053101151e-05,
|
| 142 |
+
"loss": 1.3005,
|
| 143 |
+
"step": 1900
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.2303550346972271,
|
| 147 |
+
"grad_norm": 1.321956753730774,
|
| 148 |
+
"learning_rate": 9.85579508963099e-05,
|
| 149 |
+
"loss": 1.2757,
|
| 150 |
+
"step": 2000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.24187278643208845,
|
| 154 |
+
"grad_norm": 1.3433573246002197,
|
| 155 |
+
"learning_rate": 9.841062690264163e-05,
|
| 156 |
+
"loss": 1.2155,
|
| 157 |
+
"step": 2100
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.2533905381669498,
|
| 161 |
+
"grad_norm": 1.2773957252502441,
|
| 162 |
+
"learning_rate": 9.825625998315115e-05,
|
| 163 |
+
"loss": 1.2106,
|
| 164 |
+
"step": 2200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.26490828990181114,
|
| 168 |
+
"grad_norm": 1.2487130165100098,
|
| 169 |
+
"learning_rate": 9.809487259560925e-05,
|
| 170 |
+
"loss": 1.1769,
|
| 171 |
+
"step": 2300
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2764260416366725,
|
| 175 |
+
"grad_norm": 1.2487260103225708,
|
| 176 |
+
"learning_rate": 9.792648821914587e-05,
|
| 177 |
+
"loss": 1.2285,
|
| 178 |
+
"step": 2400
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.28794379337153386,
|
| 182 |
+
"grad_norm": 2.003361940383911,
|
| 183 |
+
"learning_rate": 9.775113135083427e-05,
|
| 184 |
+
"loss": 1.1602,
|
| 185 |
+
"step": 2500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.2994615451063952,
|
| 189 |
+
"grad_norm": 1.1892813444137573,
|
| 190 |
+
"learning_rate": 9.756882750212708e-05,
|
| 191 |
+
"loss": 1.1457,
|
| 192 |
+
"step": 2600
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.3109792968412566,
|
| 196 |
+
"grad_norm": 0.9468798637390137,
|
| 197 |
+
"learning_rate": 9.737960319514486e-05,
|
| 198 |
+
"loss": 1.1555,
|
| 199 |
+
"step": 2700
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.32249704857611794,
|
| 203 |
+
"grad_norm": 1.7301925420761108,
|
| 204 |
+
"learning_rate": 9.718348595881761e-05,
|
| 205 |
+
"loss": 1.1718,
|
| 206 |
+
"step": 2800
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.3340148003109793,
|
| 210 |
+
"grad_norm": 1.5235856771469116,
|
| 211 |
+
"learning_rate": 9.698050432487968e-05,
|
| 212 |
+
"loss": 1.1165,
|
| 213 |
+
"step": 2900
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.34553255204584066,
|
| 217 |
+
"grad_norm": 1.4494585990905762,
|
| 218 |
+
"learning_rate": 9.6770687823719e-05,
|
| 219 |
+
"loss": 1.1172,
|
| 220 |
+
"step": 3000
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.357050303780702,
|
| 224 |
+
"grad_norm": 1.3340733051300049,
|
| 225 |
+
"learning_rate": 9.655406698008081e-05,
|
| 226 |
+
"loss": 1.1798,
|
| 227 |
+
"step": 3100
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.3685680555155634,
|
| 231 |
+
"grad_norm": 1.0183467864990234,
|
| 232 |
+
"learning_rate": 9.633067330862688e-05,
|
| 233 |
+
"loss": 1.1363,
|
| 234 |
+
"step": 3200
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.38008580725042473,
|
| 238 |
+
"grad_norm": 0.9531416893005371,
|
| 239 |
+
"learning_rate": 9.61005393093507e-05,
|
| 240 |
+
"loss": 1.2012,
|
| 241 |
+
"step": 3300
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.3916035589852861,
|
| 245 |
+
"grad_norm": 0.9986298680305481,
|
| 246 |
+
"learning_rate": 9.586369846284912e-05,
|
| 247 |
+
"loss": 1.1742,
|
| 248 |
+
"step": 3400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.40312131072014745,
|
| 252 |
+
"grad_norm": 1.2070988416671753,
|
| 253 |
+
"learning_rate": 9.562018522545175e-05,
|
| 254 |
+
"loss": 1.1404,
|
| 255 |
+
"step": 3500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.41463906245500876,
|
| 259 |
+
"grad_norm": 0.9883187413215637,
|
| 260 |
+
"learning_rate": 9.537003502420789e-05,
|
| 261 |
+
"loss": 1.1168,
|
| 262 |
+
"step": 3600
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.4261568141898701,
|
| 266 |
+
"grad_norm": 1.2827179431915283,
|
| 267 |
+
"learning_rate": 9.511328425173265e-05,
|
| 268 |
+
"loss": 1.167,
|
| 269 |
+
"step": 3700
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.4376745659247315,
|
| 273 |
+
"grad_norm": 1.667656421661377,
|
| 274 |
+
"learning_rate": 9.484997026091238e-05,
|
| 275 |
+
"loss": 1.073,
|
| 276 |
+
"step": 3800
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.44919231765959283,
|
| 280 |
+
"grad_norm": 1.2254990339279175,
|
| 281 |
+
"learning_rate": 9.458013135947051e-05,
|
| 282 |
+
"loss": 1.0808,
|
| 283 |
+
"step": 3900
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.4607100693944542,
|
| 287 |
+
"grad_norm": 1.6514670848846436,
|
| 288 |
+
"learning_rate": 9.430380680439435e-05,
|
| 289 |
+
"loss": 1.1928,
|
| 290 |
+
"step": 4000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.47222782112931555,
|
| 294 |
+
"grad_norm": 1.2024292945861816,
|
| 295 |
+
"learning_rate": 9.402103679622392e-05,
|
| 296 |
+
"loss": 1.0847,
|
| 297 |
+
"step": 4100
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.4837455728641769,
|
| 301 |
+
"grad_norm": 1.0424613952636719,
|
| 302 |
+
"learning_rate": 9.373186247320345e-05,
|
| 303 |
+
"loss": 1.0647,
|
| 304 |
+
"step": 4200
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.49526332459903827,
|
| 308 |
+
"grad_norm": 1.5781737565994263,
|
| 309 |
+
"learning_rate": 9.343632590529642e-05,
|
| 310 |
+
"loss": 1.1076,
|
| 311 |
+
"step": 4300
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.5067810763338996,
|
| 315 |
+
"grad_norm": 0.9091070890426636,
|
| 316 |
+
"learning_rate": 9.313447008806515e-05,
|
| 317 |
+
"loss": 1.0865,
|
| 318 |
+
"step": 4400
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.5182988280687609,
|
| 322 |
+
"grad_norm": 1.4019910097122192,
|
| 323 |
+
"learning_rate": 9.282633893641564e-05,
|
| 324 |
+
"loss": 1.074,
|
| 325 |
+
"step": 4500
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.5298165798036223,
|
| 329 |
+
"grad_norm": 1.0551828145980835,
|
| 330 |
+
"learning_rate": 9.251197727820871e-05,
|
| 331 |
+
"loss": 1.0721,
|
| 332 |
+
"step": 4600
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5413343315384836,
|
| 336 |
+
"grad_norm": 0.7485769987106323,
|
| 337 |
+
"learning_rate": 9.219143084773828e-05,
|
| 338 |
+
"loss": 1.0482,
|
| 339 |
+
"step": 4700
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.552852083273345,
|
| 343 |
+
"grad_norm": 1.1340844631195068,
|
| 344 |
+
"learning_rate": 9.186474627907784e-05,
|
| 345 |
+
"loss": 1.1191,
|
| 346 |
+
"step": 4800
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.5643698350082064,
|
| 350 |
+
"grad_norm": 1.5390311479568481,
|
| 351 |
+
"learning_rate": 9.153197109929595e-05,
|
| 352 |
+
"loss": 1.1175,
|
| 353 |
+
"step": 4900
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.5758875867430677,
|
| 357 |
+
"grad_norm": 1.3286751508712769,
|
| 358 |
+
"learning_rate": 9.119315372154189e-05,
|
| 359 |
+
"loss": 1.0089,
|
| 360 |
+
"step": 5000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.5874053384779291,
|
| 364 |
+
"grad_norm": 1.0134896039962769,
|
| 365 |
+
"learning_rate": 9.084834343800234e-05,
|
| 366 |
+
"loss": 1.0604,
|
| 367 |
+
"step": 5100
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.5989230902127904,
|
| 371 |
+
"grad_norm": 1.1101710796356201,
|
| 372 |
+
"learning_rate": 9.049759041273023e-05,
|
| 373 |
+
"loss": 1.0266,
|
| 374 |
+
"step": 5200
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.6104408419476518,
|
| 378 |
+
"grad_norm": 0.9307985901832581,
|
| 379 |
+
"learning_rate": 9.014094567434668e-05,
|
| 380 |
+
"loss": 1.042,
|
| 381 |
+
"step": 5300
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.6219585936825132,
|
| 385 |
+
"grad_norm": 1.295506238937378,
|
| 386 |
+
"learning_rate": 8.977846110861722e-05,
|
| 387 |
+
"loss": 1.102,
|
| 388 |
+
"step": 5400
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.6334763454173745,
|
| 392 |
+
"grad_norm": 1.0131995677947998,
|
| 393 |
+
"learning_rate": 8.941018945090334e-05,
|
| 394 |
+
"loss": 1.0017,
|
| 395 |
+
"step": 5500
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.6449940971522359,
|
| 399 |
+
"grad_norm": 0.9528135061264038,
|
| 400 |
+
"learning_rate": 8.90361842784903e-05,
|
| 401 |
+
"loss": 1.0597,
|
| 402 |
+
"step": 5600
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.6565118488870972,
|
| 406 |
+
"grad_norm": 0.8767136335372925,
|
| 407 |
+
"learning_rate": 8.86565000027926e-05,
|
| 408 |
+
"loss": 1.0354,
|
| 409 |
+
"step": 5700
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.6680296006219586,
|
| 413 |
+
"grad_norm": 0.7215752601623535,
|
| 414 |
+
"learning_rate": 8.8271191861438e-05,
|
| 415 |
+
"loss": 1.0266,
|
| 416 |
+
"step": 5800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.67954735235682,
|
| 420 |
+
"grad_norm": 0.9168160557746887,
|
| 421 |
+
"learning_rate": 8.788031591023145e-05,
|
| 422 |
+
"loss": 0.9974,
|
| 423 |
+
"step": 5900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.6910651040916813,
|
| 427 |
+
"grad_norm": 0.9805298447608948,
|
| 428 |
+
"learning_rate": 8.748392901499985e-05,
|
| 429 |
+
"loss": 1.0205,
|
| 430 |
+
"step": 6000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.7025828558265427,
|
| 434 |
+
"grad_norm": 0.9025911092758179,
|
| 435 |
+
"learning_rate": 8.708208884331907e-05,
|
| 436 |
+
"loss": 1.0282,
|
| 437 |
+
"step": 6100
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.714100607561404,
|
| 441 |
+
"grad_norm": 0.7643795609474182,
|
| 442 |
+
"learning_rate": 8.667485385612433e-05,
|
| 443 |
+
"loss": 1.0281,
|
| 444 |
+
"step": 6200
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.7256183592962654,
|
| 448 |
+
"grad_norm": 1.015793800354004,
|
| 449 |
+
"learning_rate": 8.6262283299205e-05,
|
| 450 |
+
"loss": 1.0153,
|
| 451 |
+
"step": 6300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.7371361110311268,
|
| 455 |
+
"grad_norm": 0.9706717133522034,
|
| 456 |
+
"learning_rate": 8.58444371945855e-05,
|
| 457 |
+
"loss": 0.9879,
|
| 458 |
+
"step": 6400
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.7486538627659881,
|
| 462 |
+
"grad_norm": 0.8967713117599487,
|
| 463 |
+
"learning_rate": 8.5421376331793e-05,
|
| 464 |
+
"loss": 1.044,
|
| 465 |
+
"step": 6500
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.7601716145008495,
|
| 469 |
+
"grad_norm": 1.0314196348190308,
|
| 470 |
+
"learning_rate": 8.499316225901364e-05,
|
| 471 |
+
"loss": 1.0163,
|
| 472 |
+
"step": 6600
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.7716893662357108,
|
| 476 |
+
"grad_norm": 0.9607068300247192,
|
| 477 |
+
"learning_rate": 8.455985727413825e-05,
|
| 478 |
+
"loss": 1.0047,
|
| 479 |
+
"step": 6700
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.7832071179705722,
|
| 483 |
+
"grad_norm": 0.8781064748764038,
|
| 484 |
+
"learning_rate": 8.41215244156991e-05,
|
| 485 |
+
"loss": 0.9683,
|
| 486 |
+
"step": 6800
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.7947248697054335,
|
| 490 |
+
"grad_norm": 0.7523179054260254,
|
| 491 |
+
"learning_rate": 8.367822745369891e-05,
|
| 492 |
+
"loss": 0.9814,
|
| 493 |
+
"step": 6900
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.8062426214402949,
|
| 497 |
+
"grad_norm": 0.7084505558013916,
|
| 498 |
+
"learning_rate": 8.323003088033327e-05,
|
| 499 |
+
"loss": 1.0103,
|
| 500 |
+
"step": 7000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.8177603731751563,
|
| 504 |
+
"grad_norm": 0.8509836792945862,
|
| 505 |
+
"learning_rate": 8.277699990060826e-05,
|
| 506 |
+
"loss": 0.947,
|
| 507 |
+
"step": 7100
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.8292781249100175,
|
| 511 |
+
"grad_norm": 0.9924744963645935,
|
| 512 |
+
"learning_rate": 8.231920042285408e-05,
|
| 513 |
+
"loss": 0.9985,
|
| 514 |
+
"step": 7200
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.8407958766448789,
|
| 518 |
+
"grad_norm": 0.9982401132583618,
|
| 519 |
+
"learning_rate": 8.185669904913669e-05,
|
| 520 |
+
"loss": 1.004,
|
| 521 |
+
"step": 7300
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.8523136283797402,
|
| 525 |
+
"grad_norm": 0.9341840744018555,
|
| 526 |
+
"learning_rate": 8.138956306556813e-05,
|
| 527 |
+
"loss": 1.005,
|
| 528 |
+
"step": 7400
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.8638313801146016,
|
| 532 |
+
"grad_norm": 1.4367643594741821,
|
| 533 |
+
"learning_rate": 8.09178604325177e-05,
|
| 534 |
+
"loss": 0.9949,
|
| 535 |
+
"step": 7500
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.875349131849463,
|
| 539 |
+
"grad_norm": 0.7783756256103516,
|
| 540 |
+
"learning_rate": 8.044165977472476e-05,
|
| 541 |
+
"loss": 0.9786,
|
| 542 |
+
"step": 7600
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.8868668835843243,
|
| 546 |
+
"grad_norm": 1.033542513847351,
|
| 547 |
+
"learning_rate": 7.99610303713151e-05,
|
| 548 |
+
"loss": 0.9979,
|
| 549 |
+
"step": 7700
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.8983846353191857,
|
| 553 |
+
"grad_norm": 0.7946944832801819,
|
| 554 |
+
"learning_rate": 7.947604214572189e-05,
|
| 555 |
+
"loss": 0.9736,
|
| 556 |
+
"step": 7800
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.909902387054047,
|
| 560 |
+
"grad_norm": 0.8648232221603394,
|
| 561 |
+
"learning_rate": 7.89867656555131e-05,
|
| 562 |
+
"loss": 0.9916,
|
| 563 |
+
"step": 7900
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.9214201387889084,
|
| 567 |
+
"grad_norm": 1.1316163539886475,
|
| 568 |
+
"learning_rate": 7.849327208212653e-05,
|
| 569 |
+
"loss": 1.0083,
|
| 570 |
+
"step": 8000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.9329378905237697,
|
| 574 |
+
"grad_norm": 0.731547474861145,
|
| 575 |
+
"learning_rate": 7.799563322051407e-05,
|
| 576 |
+
"loss": 0.9868,
|
| 577 |
+
"step": 8100
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.9444556422586311,
|
| 581 |
+
"grad_norm": 0.7558659315109253,
|
| 582 |
+
"learning_rate": 7.749392146869688e-05,
|
| 583 |
+
"loss": 0.9378,
|
| 584 |
+
"step": 8200
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.9559733939934925,
|
| 588 |
+
"grad_norm": 1.1830765008926392,
|
| 589 |
+
"learning_rate": 7.698820981723252e-05,
|
| 590 |
+
"loss": 0.947,
|
| 591 |
+
"step": 8300
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.9674911457283538,
|
| 595 |
+
"grad_norm": 1.1989309787750244,
|
| 596 |
+
"learning_rate": 7.647857183859626e-05,
|
| 597 |
+
"loss": 0.9973,
|
| 598 |
+
"step": 8400
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.9790088974632152,
|
| 602 |
+
"grad_norm": 0.7769212126731873,
|
| 603 |
+
"learning_rate": 7.596508167647738e-05,
|
| 604 |
+
"loss": 0.983,
|
| 605 |
+
"step": 8500
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.9905266491980765,
|
| 609 |
+
"grad_norm": 0.8555830121040344,
|
| 610 |
+
"learning_rate": 7.544781403499261e-05,
|
| 611 |
+
"loss": 0.9472,
|
| 612 |
+
"step": 8600
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.0019580177949263,
|
| 616 |
+
"grad_norm": 0.9663079380989075,
|
| 617 |
+
"learning_rate": 7.492684416781785e-05,
|
| 618 |
+
"loss": 0.9254,
|
| 619 |
+
"step": 8700
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.0134757695297878,
|
| 623 |
+
"grad_norm": 1.0730838775634766,
|
| 624 |
+
"learning_rate": 7.44022478672402e-05,
|
| 625 |
+
"loss": 0.7473,
|
| 626 |
+
"step": 8800
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.024993521264649,
|
| 630 |
+
"grad_norm": 1.2075188159942627,
|
| 631 |
+
"learning_rate": 7.38741014531313e-05,
|
| 632 |
+
"loss": 0.7745,
|
| 633 |
+
"step": 8900
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.0365112729995105,
|
| 637 |
+
"grad_norm": 0.724323034286499,
|
| 638 |
+
"learning_rate": 7.334248176184411e-05,
|
| 639 |
+
"loss": 0.7592,
|
| 640 |
+
"step": 9000
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.0480290247343718,
|
| 644 |
+
"grad_norm": 0.9635033011436462,
|
| 645 |
+
"learning_rate": 7.280746613503463e-05,
|
| 646 |
+
"loss": 0.7708,
|
| 647 |
+
"step": 9100
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.0595467764692332,
|
| 651 |
+
"grad_norm": 1.1375808715820312,
|
| 652 |
+
"learning_rate": 7.226913240840992e-05,
|
| 653 |
+
"loss": 0.7739,
|
| 654 |
+
"step": 9200
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.0710645282040945,
|
| 658 |
+
"grad_norm": 1.0138040781021118,
|
| 659 |
+
"learning_rate": 7.172755890040428e-05,
|
| 660 |
+
"loss": 0.7582,
|
| 661 |
+
"step": 9300
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.082582279938956,
|
| 665 |
+
"grad_norm": 0.9169549345970154,
|
| 666 |
+
"learning_rate": 7.118282440078535e-05,
|
| 667 |
+
"loss": 0.7488,
|
| 668 |
+
"step": 9400
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.0941000316738172,
|
| 672 |
+
"grad_norm": 0.8004472851753235,
|
| 673 |
+
"learning_rate": 7.063500815919148e-05,
|
| 674 |
+
"loss": 0.7401,
|
| 675 |
+
"step": 9500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.1056177834086787,
|
| 679 |
+
"grad_norm": 1.0395482778549194,
|
| 680 |
+
"learning_rate": 7.008418987360222e-05,
|
| 681 |
+
"loss": 0.7769,
|
| 682 |
+
"step": 9600
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.11713553514354,
|
| 686 |
+
"grad_norm": 1.0113108158111572,
|
| 687 |
+
"learning_rate": 6.953044967874367e-05,
|
| 688 |
+
"loss": 0.8,
|
| 689 |
+
"step": 9700
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.1286532868784014,
|
| 693 |
+
"grad_norm": 0.8164701461791992,
|
| 694 |
+
"learning_rate": 6.897386813443019e-05,
|
| 695 |
+
"loss": 0.7473,
|
| 696 |
+
"step": 9800
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.1401710386132626,
|
| 700 |
+
"grad_norm": 0.7086979150772095,
|
| 701 |
+
"learning_rate": 6.84145262138444e-05,
|
| 702 |
+
"loss": 0.7732,
|
| 703 |
+
"step": 9900
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.151688790348124,
|
| 707 |
+
"grad_norm": 1.1815799474716187,
|
| 708 |
+
"learning_rate": 6.785250529175679e-05,
|
| 709 |
+
"loss": 0.7656,
|
| 710 |
+
"step": 10000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.151688790348124,
|
| 714 |
+
"eval_gen_len": 0.0,
|
| 715 |
+
"eval_loss": 0.12472382932901382,
|
| 716 |
+
"eval_rouge1": 0.0,
|
| 717 |
+
"eval_rouge2": 0.0,
|
| 718 |
+
"eval_rougeL": 0.0,
|
| 719 |
+
"eval_rougeLsum": 0.0,
|
| 720 |
+
"eval_runtime": 474.3304,
|
| 721 |
+
"eval_samples_per_second": 4.025,
|
| 722 |
+
"eval_steps_per_second": 0.673,
|
| 723 |
+
"step": 10000
|
| 724 |
+
}
|
| 725 |
+
],
|
| 726 |
+
"logging_steps": 100,
|
| 727 |
+
"max_steps": 26046,
|
| 728 |
+
"num_input_tokens_seen": 0,
|
| 729 |
+
"num_train_epochs": 3,
|
| 730 |
+
"save_steps": 10000,
|
| 731 |
+
"stateful_callbacks": {
|
| 732 |
+
"EarlyStoppingCallback": {
|
| 733 |
+
"args": {
|
| 734 |
+
"early_stopping_patience": 3,
|
| 735 |
+
"early_stopping_threshold": 0.0
|
| 736 |
+
},
|
| 737 |
+
"attributes": {
|
| 738 |
+
"early_stopping_patience_counter": 0
|
| 739 |
+
}
|
| 740 |
+
},
|
| 741 |
+
"TrainerControl": {
|
| 742 |
+
"args": {
|
| 743 |
+
"should_epoch_stop": false,
|
| 744 |
+
"should_evaluate": false,
|
| 745 |
+
"should_log": false,
|
| 746 |
+
"should_save": true,
|
| 747 |
+
"should_training_stop": false
|
| 748 |
+
},
|
| 749 |
+
"attributes": {}
|
| 750 |
+
}
|
| 751 |
+
},
|
| 752 |
+
"total_flos": 6.24479560268356e+18,
|
| 753 |
+
"train_batch_size": 6,
|
| 754 |
+
"trial_name": null,
|
| 755 |
+
"trial_params": null
|
| 756 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e4bede8800d405d3ba26c59be83219d2f9a20f60d7968e24f81627965363e0af
|
| 3 |
+
size 5496
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|