Add step-10000 checkpoint
Browse files- config.json +192 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- text_renderer_config.json +225 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1120 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,192 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"PIXELSumModel"
|
| 4 |
+
],
|
| 5 |
+
"decoder": {
|
| 6 |
+
"_name_or_path": "gpt2",
|
| 7 |
+
"activation_function": "gelu_new",
|
| 8 |
+
"add_cross_attention": true,
|
| 9 |
+
"architectures": [
|
| 10 |
+
"GPT2LMHeadModel"
|
| 11 |
+
],
|
| 12 |
+
"attn_pdrop": 0.1,
|
| 13 |
+
"bad_words_ids": null,
|
| 14 |
+
"bos_token_id": 50256,
|
| 15 |
+
"chunk_size_feed_forward": 0,
|
| 16 |
+
"cross_attention_hidden_size": null,
|
| 17 |
+
"cross_attention_reduce_factor": 1,
|
| 18 |
+
"decoder_start_token_id": null,
|
| 19 |
+
"diversity_penalty": 0.0,
|
| 20 |
+
"do_sample": false,
|
| 21 |
+
"early_stopping": false,
|
| 22 |
+
"embd_pdrop": 0.1,
|
| 23 |
+
"encoder_hidden_size": 768,
|
| 24 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 25 |
+
"eos_token_id": 50256,
|
| 26 |
+
"exponential_decay_length_penalty": null,
|
| 27 |
+
"finetuning_task": null,
|
| 28 |
+
"forced_bos_token_id": null,
|
| 29 |
+
"forced_eos_token_id": null,
|
| 30 |
+
"id2label": {
|
| 31 |
+
"0": "LABEL_0",
|
| 32 |
+
"1": "LABEL_1"
|
| 33 |
+
},
|
| 34 |
+
"initializer_range": 0.02,
|
| 35 |
+
"is_decoder": true,
|
| 36 |
+
"is_encoder_decoder": false,
|
| 37 |
+
"label2id": {
|
| 38 |
+
"LABEL_0": 0,
|
| 39 |
+
"LABEL_1": 1
|
| 40 |
+
},
|
| 41 |
+
"layer_norm_epsilon": 1e-05,
|
| 42 |
+
"length_penalty": 1.0,
|
| 43 |
+
"max_length": 20,
|
| 44 |
+
"min_length": 0,
|
| 45 |
+
"model_type": "gpt2",
|
| 46 |
+
"n_ctx": 1024,
|
| 47 |
+
"n_embd": 768,
|
| 48 |
+
"n_head": 12,
|
| 49 |
+
"n_inner": null,
|
| 50 |
+
"n_layer": 12,
|
| 51 |
+
"n_positions": 1024,
|
| 52 |
+
"no_repeat_ngram_size": 0,
|
| 53 |
+
"num_beam_groups": 1,
|
| 54 |
+
"num_beams": 1,
|
| 55 |
+
"num_return_sequences": 1,
|
| 56 |
+
"output_attentions": false,
|
| 57 |
+
"output_hidden_states": false,
|
| 58 |
+
"output_scores": false,
|
| 59 |
+
"pad_token_id": null,
|
| 60 |
+
"prefix": null,
|
| 61 |
+
"problem_type": null,
|
| 62 |
+
"pruned_heads": {},
|
| 63 |
+
"remove_invalid_values": false,
|
| 64 |
+
"reorder_and_upcast_attn": false,
|
| 65 |
+
"repetition_penalty": 1.0,
|
| 66 |
+
"resid_pdrop": 0.1,
|
| 67 |
+
"return_dict": true,
|
| 68 |
+
"return_dict_in_generate": false,
|
| 69 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 70 |
+
"scale_attn_weights": true,
|
| 71 |
+
"sep_token_id": null,
|
| 72 |
+
"summary_activation": null,
|
| 73 |
+
"summary_first_dropout": 0.1,
|
| 74 |
+
"summary_proj_to_labels": true,
|
| 75 |
+
"summary_type": "cls_index",
|
| 76 |
+
"summary_use_proj": true,
|
| 77 |
+
"task_specific_params": {
|
| 78 |
+
"text-generation": {
|
| 79 |
+
"do_sample": true,
|
| 80 |
+
"max_length": 64
|
| 81 |
+
}
|
| 82 |
+
},
|
| 83 |
+
"temperature": 1.0,
|
| 84 |
+
"tie_encoder_decoder": false,
|
| 85 |
+
"tie_word_embeddings": true,
|
| 86 |
+
"tokenizer_class": null,
|
| 87 |
+
"top_k": 50,
|
| 88 |
+
"top_p": 1.0,
|
| 89 |
+
"torch_dtype": null,
|
| 90 |
+
"torchscript": false,
|
| 91 |
+
"transformers_version": "4.19.0",
|
| 92 |
+
"typical_p": 1.0,
|
| 93 |
+
"use_bfloat16": false,
|
| 94 |
+
"use_cache": true,
|
| 95 |
+
"vocab_size": 50257
|
| 96 |
+
},
|
| 97 |
+
"decoder_start_token_id": 50256,
|
| 98 |
+
"encoder": {
|
| 99 |
+
"_name_or_path": "Team-PIXEL/pixel-base",
|
| 100 |
+
"add_cross_attention": false,
|
| 101 |
+
"architectures": [
|
| 102 |
+
"PIXELForPreTraining"
|
| 103 |
+
],
|
| 104 |
+
"attention_probs_dropout_prob": 0.1,
|
| 105 |
+
"bad_words_ids": null,
|
| 106 |
+
"bos_token_id": null,
|
| 107 |
+
"chunk_size_feed_forward": 0,
|
| 108 |
+
"cross_attention_hidden_size": null,
|
| 109 |
+
"decoder_hidden_size": 512,
|
| 110 |
+
"decoder_intermediate_size": 2048,
|
| 111 |
+
"decoder_num_attention_heads": 16,
|
| 112 |
+
"decoder_num_hidden_layers": 8,
|
| 113 |
+
"decoder_start_token_id": null,
|
| 114 |
+
"diversity_penalty": 0.0,
|
| 115 |
+
"do_eval": true,
|
| 116 |
+
"do_sample": false,
|
| 117 |
+
"early_stopping": false,
|
| 118 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 119 |
+
"eos_token_id": null,
|
| 120 |
+
"exponential_decay_length_penalty": null,
|
| 121 |
+
"finetuning_task": null,
|
| 122 |
+
"forced_bos_token_id": null,
|
| 123 |
+
"forced_eos_token_id": null,
|
| 124 |
+
"hidden_act": "gelu",
|
| 125 |
+
"hidden_dropout_prob": 0.1,
|
| 126 |
+
"hidden_size": 768,
|
| 127 |
+
"id2label": {
|
| 128 |
+
"0": "LABEL_0",
|
| 129 |
+
"1": "LABEL_1"
|
| 130 |
+
},
|
| 131 |
+
"image_size": [
|
| 132 |
+
16,
|
| 133 |
+
8464
|
| 134 |
+
],
|
| 135 |
+
"initializer_range": 0.02,
|
| 136 |
+
"intermediate_size": 3072,
|
| 137 |
+
"is_decoder": false,
|
| 138 |
+
"is_encoder_decoder": false,
|
| 139 |
+
"label2id": {
|
| 140 |
+
"LABEL_0": 0,
|
| 141 |
+
"LABEL_1": 1
|
| 142 |
+
},
|
| 143 |
+
"layer_norm_eps": 1e-12,
|
| 144 |
+
"length_penalty": 1.0,
|
| 145 |
+
"mask_ratio": 0.25,
|
| 146 |
+
"max_length": 20,
|
| 147 |
+
"min_length": 0,
|
| 148 |
+
"model_type": "pixel",
|
| 149 |
+
"no_repeat_ngram_size": 0,
|
| 150 |
+
"norm_pix_loss": true,
|
| 151 |
+
"num_attention_heads": 12,
|
| 152 |
+
"num_beam_groups": 1,
|
| 153 |
+
"num_beams": 1,
|
| 154 |
+
"num_channels": 3,
|
| 155 |
+
"num_hidden_layers": 12,
|
| 156 |
+
"num_return_sequences": 1,
|
| 157 |
+
"output_attentions": false,
|
| 158 |
+
"output_hidden_states": false,
|
| 159 |
+
"output_scores": false,
|
| 160 |
+
"pad_token_id": null,
|
| 161 |
+
"patch_size": 16,
|
| 162 |
+
"prefix": null,
|
| 163 |
+
"problem_type": null,
|
| 164 |
+
"pruned_heads": {},
|
| 165 |
+
"qkv_bias": true,
|
| 166 |
+
"remove_invalid_values": false,
|
| 167 |
+
"repetition_penalty": 1.0,
|
| 168 |
+
"return_dict": true,
|
| 169 |
+
"return_dict_in_generate": false,
|
| 170 |
+
"sep_token_id": null,
|
| 171 |
+
"task_specific_params": null,
|
| 172 |
+
"temperature": 1.0,
|
| 173 |
+
"tie_encoder_decoder": false,
|
| 174 |
+
"tie_word_embeddings": true,
|
| 175 |
+
"tokenizer_class": null,
|
| 176 |
+
"top_k": 50,
|
| 177 |
+
"top_p": 1.0,
|
| 178 |
+
"torch_dtype": "float32",
|
| 179 |
+
"torchscript": false,
|
| 180 |
+
"transformers_version": "4.19.0",
|
| 181 |
+
"typical_p": 1.0,
|
| 182 |
+
"use_bfloat16": false
|
| 183 |
+
},
|
| 184 |
+
"eos_token_id": 50256,
|
| 185 |
+
"is_encoder_decoder": true,
|
| 186 |
+
"model_type": "pixelsum",
|
| 187 |
+
"pad_token_id": 50256,
|
| 188 |
+
"tie_word_embeddings": false,
|
| 189 |
+
"torch_dtype": "float32",
|
| 190 |
+
"transformers_version": null,
|
| 191 |
+
"vocab_size": 50257
|
| 192 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c0c2461a01d8acf726a91c0121ed4716c53e02dec1dff23ef41f70d85d8005b
|
| 3 |
+
size 980777906
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
text_renderer_config.json
ADDED
|
@@ -0,0 +1,225 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"background_color": "white",
|
| 3 |
+
"dpi": 120,
|
| 4 |
+
"font_color": "black",
|
| 5 |
+
"font_file": "49e6dc219d1a1a1c9236acaf05a48b542002016a6dc877ee72baab085a84257b.3f28e7f4b38e1efe1b6da4a3732404c19d4c6a614ff32dce90a251e293d4ce58",
|
| 6 |
+
"font_size": 8,
|
| 7 |
+
"fonts_list": [
|
| 8 |
+
"Apple Color Emoji",
|
| 9 |
+
"DejaVu Sans",
|
| 10 |
+
"DejaVu Sans Mono",
|
| 11 |
+
"DejaVu Serif",
|
| 12 |
+
"Go Noto Current",
|
| 13 |
+
"Inconsolata",
|
| 14 |
+
"Monospace",
|
| 15 |
+
"Noto Fangsong KSS Rotated",
|
| 16 |
+
"Noto Fangsong KSS Vertical",
|
| 17 |
+
"Noto Kufi Arabic",
|
| 18 |
+
"Noto Music",
|
| 19 |
+
"Noto Naskh Arabic",
|
| 20 |
+
"Noto Naskh Arabic UI",
|
| 21 |
+
"Noto Nastaliq Urdu",
|
| 22 |
+
"Noto Rashi Hebrew",
|
| 23 |
+
"Noto Sans",
|
| 24 |
+
"Noto Sans Adlam",
|
| 25 |
+
"Noto Sans Adlam Unjoined",
|
| 26 |
+
"Noto Sans Anatolian Hieroglyphs",
|
| 27 |
+
"Noto Sans Arabic",
|
| 28 |
+
"Noto Sans Arabic UI",
|
| 29 |
+
"Noto Sans Armenian",
|
| 30 |
+
"Noto Sans Avestan",
|
| 31 |
+
"Noto Sans Balinese",
|
| 32 |
+
"Noto Sans Bamum",
|
| 33 |
+
"Noto Sans Bassa Vah",
|
| 34 |
+
"Noto Sans Batak",
|
| 35 |
+
"Noto Sans Bhaiksuki",
|
| 36 |
+
"Noto Sans Brahmi",
|
| 37 |
+
"Noto Sans Buginese",
|
| 38 |
+
"Noto Sans Buhid",
|
| 39 |
+
"Noto Sans CJK HK",
|
| 40 |
+
"Noto Sans CJK JP",
|
| 41 |
+
"Noto Sans CJK KR",
|
| 42 |
+
"Noto Sans CJK SC",
|
| 43 |
+
"Noto Sans CJK TC",
|
| 44 |
+
"Noto Sans Canadian Aboriginal",
|
| 45 |
+
"Noto Sans Carian",
|
| 46 |
+
"Noto Sans Caucasian Albanian",
|
| 47 |
+
"Noto Sans Chakma",
|
| 48 |
+
"Noto Sans Cham",
|
| 49 |
+
"Noto Sans Cherokee",
|
| 50 |
+
"Noto Sans Chorasmian",
|
| 51 |
+
"Noto Sans Coptic",
|
| 52 |
+
"Noto Sans Cuneiform",
|
| 53 |
+
"Noto Sans Cypriot",
|
| 54 |
+
"Noto Sans Cypro Minoan",
|
| 55 |
+
"Noto Sans Deseret",
|
| 56 |
+
"Noto Sans Devanagari",
|
| 57 |
+
"Noto Sans Devanagari UI",
|
| 58 |
+
"Noto Sans Duployan",
|
| 59 |
+
"Noto Sans Egyptian Hieroglyphs",
|
| 60 |
+
"Noto Sans Elbasan",
|
| 61 |
+
"Noto Sans Elymaic",
|
| 62 |
+
"Noto Sans Ethiopic",
|
| 63 |
+
"Noto Sans Georgian",
|
| 64 |
+
"Noto Sans Glagolitic",
|
| 65 |
+
"Noto Sans Gothic",
|
| 66 |
+
"Noto Sans Grantha",
|
| 67 |
+
"Noto Sans Gujarati",
|
| 68 |
+
"Noto Sans Gujarati UI",
|
| 69 |
+
"Noto Sans Gunjala Gondi",
|
| 70 |
+
"Noto Sans Gurmukhi",
|
| 71 |
+
"Noto Sans Gurmukhi UI",
|
| 72 |
+
"Noto Sans Hanifi Rohingya",
|
| 73 |
+
"Noto Sans Hanunoo",
|
| 74 |
+
"Noto Sans Hatran",
|
| 75 |
+
"Noto Sans Hebrew",
|
| 76 |
+
"Noto Sans Hebrew Droid",
|
| 77 |
+
"Noto Sans Imperial Aramaic",
|
| 78 |
+
"Noto Sans Indic Siyaq Numbers",
|
| 79 |
+
"Noto Sans Inscriptional Pahlavi",
|
| 80 |
+
"Noto Sans Inscriptional Parthian",
|
| 81 |
+
"Noto Sans Javanese",
|
| 82 |
+
"Noto Sans Kaithi",
|
| 83 |
+
"Noto Sans Kannada",
|
| 84 |
+
"Noto Sans Kannada UI",
|
| 85 |
+
"Noto Sans Kawi",
|
| 86 |
+
"Noto Sans Kayah Li",
|
| 87 |
+
"Noto Sans Kharoshthi",
|
| 88 |
+
"Noto Sans Khmer",
|
| 89 |
+
"Noto Sans Khmer UI",
|
| 90 |
+
"Noto Sans Khudawadi",
|
| 91 |
+
"Noto Sans Lao",
|
| 92 |
+
"Noto Sans Lao Looped",
|
| 93 |
+
"Noto Sans Lao Looped UI",
|
| 94 |
+
"Noto Sans Lao UI",
|
| 95 |
+
"Noto Sans Lepcha",
|
| 96 |
+
"Noto Sans Limbu",
|
| 97 |
+
"Noto Sans Linear A",
|
| 98 |
+
"Noto Sans Linear B",
|
| 99 |
+
"Noto Sans Lisu",
|
| 100 |
+
"Noto Sans Lycian",
|
| 101 |
+
"Noto Sans Lydian",
|
| 102 |
+
"Noto Sans Mahajani",
|
| 103 |
+
"Noto Sans Malayalam",
|
| 104 |
+
"Noto Sans Malayalam UI",
|
| 105 |
+
"Noto Sans Mandaic",
|
| 106 |
+
"Noto Sans Manichaean",
|
| 107 |
+
"Noto Sans Marchen",
|
| 108 |
+
"Noto Sans Masaram Gondi",
|
| 109 |
+
"Noto Sans Math",
|
| 110 |
+
"Noto Sans Mayan Numerals",
|
| 111 |
+
"Noto Sans Medefaidrin",
|
| 112 |
+
"Noto Sans Meetei Mayek",
|
| 113 |
+
"Noto Sans Mende Kikakui",
|
| 114 |
+
"Noto Sans Meroitic",
|
| 115 |
+
"Noto Sans Miao",
|
| 116 |
+
"Noto Sans Modi",
|
| 117 |
+
"Noto Sans Mongolian",
|
| 118 |
+
"Noto Sans Mono",
|
| 119 |
+
"Noto Sans Mro",
|
| 120 |
+
"Noto Sans Multani",
|
| 121 |
+
"Noto Sans Myanmar",
|
| 122 |
+
"Noto Sans Myanmar UI",
|
| 123 |
+
"Noto Sans NKo",
|
| 124 |
+
"Noto Sans NKo Unjoined",
|
| 125 |
+
"Noto Sans Nabataean",
|
| 126 |
+
"Noto Sans Nag Mundari",
|
| 127 |
+
"Noto Sans Nandinagari",
|
| 128 |
+
"Noto Sans New Tai Lue",
|
| 129 |
+
"Noto Sans Newa",
|
| 130 |
+
"Noto Sans Nushu",
|
| 131 |
+
"Noto Sans Ogham",
|
| 132 |
+
"Noto Sans Ol Chiki",
|
| 133 |
+
"Noto Sans Old Hungarian",
|
| 134 |
+
"Noto Sans Old Italic",
|
| 135 |
+
"Noto Sans Old North Arabian",
|
| 136 |
+
"Noto Sans Old Permic",
|
| 137 |
+
"Noto Sans Old Persian",
|
| 138 |
+
"Noto Sans Old Sogdian",
|
| 139 |
+
"Noto Sans Old South Arabian",
|
| 140 |
+
"Noto Sans Old Turkic",
|
| 141 |
+
"Noto Sans Oriya",
|
| 142 |
+
"Noto Sans Osage",
|
| 143 |
+
"Noto Sans Osmanya",
|
| 144 |
+
"Noto Sans Pahawh Hmong",
|
| 145 |
+
"Noto Sans Palmyrene",
|
| 146 |
+
"Noto Sans Pau Cin Hau",
|
| 147 |
+
"Noto Sans PhagsPa",
|
| 148 |
+
"Noto Sans Phoenician",
|
| 149 |
+
"Noto Sans Psalter Pahlavi",
|
| 150 |
+
"Noto Sans Rejang",
|
| 151 |
+
"Noto Sans Runic",
|
| 152 |
+
"Noto Sans Samaritan",
|
| 153 |
+
"Noto Sans Saurashtra",
|
| 154 |
+
"Noto Sans Sharada",
|
| 155 |
+
"Noto Sans Shavian",
|
| 156 |
+
"Noto Sans Siddham",
|
| 157 |
+
"Noto Sans SignWriting",
|
| 158 |
+
"Noto Sans Sinhala",
|
| 159 |
+
"Noto Sans Sinhala UI",
|
| 160 |
+
"Noto Sans Sogdian",
|
| 161 |
+
"Noto Sans Sora Sompeng",
|
| 162 |
+
"Noto Sans Soyombo",
|
| 163 |
+
"Noto Sans Sundanese",
|
| 164 |
+
"Noto Sans Sunuwar",
|
| 165 |
+
"Noto Sans Syloti Nagri",
|
| 166 |
+
"Noto Sans Symbols",
|
| 167 |
+
"Noto Sans Symbols 2",
|
| 168 |
+
"Noto Sans Syriac",
|
| 169 |
+
"Noto Sans Syriac Eastern",
|
| 170 |
+
"Noto Sans Syriac Western",
|
| 171 |
+
"Noto Sans Tagalog",
|
| 172 |
+
"Noto Sans Tagbanwa",
|
| 173 |
+
"Noto Sans Tai Le",
|
| 174 |
+
"Noto Sans Tai Tham",
|
| 175 |
+
"Noto Sans Tai Viet",
|
| 176 |
+
"Noto Sans Takri",
|
| 177 |
+
"Noto Sans Tamil",
|
| 178 |
+
"Noto Sans Tamil Supplement",
|
| 179 |
+
"Noto Sans Tamil UI",
|
| 180 |
+
"Noto Sans Tangsa",
|
| 181 |
+
"Noto Sans Telugu",
|
| 182 |
+
"Noto Sans Telugu UI",
|
| 183 |
+
"Noto Sans Test",
|
| 184 |
+
"Noto Sans Thaana",
|
| 185 |
+
"Noto Sans Thai",
|
| 186 |
+
"Noto Sans Thai Looped",
|
| 187 |
+
"Noto Sans Thai Looped UI",
|
| 188 |
+
"Noto Sans Thai UI",
|
| 189 |
+
"Noto Sans Tifinagh",
|
| 190 |
+
"Noto Sans Tifinagh APT",
|
| 191 |
+
"Noto Sans Tifinagh Adrar",
|
| 192 |
+
"Noto Sans Tifinagh Agraw Imazighen",
|
| 193 |
+
"Noto Sans Tifinagh Ahaggar",
|
| 194 |
+
"Noto Sans Tifinagh Air",
|
| 195 |
+
"Noto Sans Tifinagh Azawagh",
|
| 196 |
+
"Noto Sans Tifinagh Ghat",
|
| 197 |
+
"Noto Sans Tifinagh Hawad",
|
| 198 |
+
"Noto Sans Tifinagh Rhissa Ixa",
|
| 199 |
+
"Noto Sans Tifinagh SIL",
|
| 200 |
+
"Noto Sans Tifinagh Tawellemmet",
|
| 201 |
+
"Noto Sans Tirhuta",
|
| 202 |
+
"Noto Sans Ugaritic",
|
| 203 |
+
"Noto Sans Vai",
|
| 204 |
+
"Noto Sans Vithkuqi",
|
| 205 |
+
"Noto Sans Wancho",
|
| 206 |
+
"Noto Sans Warang Citi",
|
| 207 |
+
"Noto Sans Yi",
|
| 208 |
+
"Noto Sans Zanabazar Square",
|
| 209 |
+
"Noto Traditional Nushu",
|
| 210 |
+
"Noto Znamenny Musical Notation",
|
| 211 |
+
"NotoSansOldHungarianUI",
|
| 212 |
+
"Sans",
|
| 213 |
+
"Serif",
|
| 214 |
+
"Source Code Pro",
|
| 215 |
+
"System-ui",
|
| 216 |
+
"Ubuntu",
|
| 217 |
+
"Ubuntu Condensed",
|
| 218 |
+
"Ubuntu Mono"
|
| 219 |
+
],
|
| 220 |
+
"max_seq_length": 529,
|
| 221 |
+
"pad_size": 3,
|
| 222 |
+
"pixels_per_patch": 16,
|
| 223 |
+
"rgb": false,
|
| 224 |
+
"text_renderer_type": "PangoCairoTextRenderer"
|
| 225 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-base-uncased", "tokenizer_class": "BertTokenizer"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.98361849433367,
|
| 3 |
+
"best_model_checkpoint": "experiments/translation/gpt2/2025-06-24_10-08-42/checkpoint-8000",
|
| 4 |
+
"epoch": 6.4617558022199795,
|
| 5 |
+
"global_step": 8000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 4.8e-06,
|
| 13 |
+
"loss": 5.4804,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.08,
|
| 18 |
+
"learning_rate": 9.799999999999998e-06,
|
| 19 |
+
"loss": 4.9145,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.12,
|
| 24 |
+
"learning_rate": 1.4799999999999999e-05,
|
| 25 |
+
"loss": 4.7992,
|
| 26 |
+
"step": 150
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.16,
|
| 30 |
+
"learning_rate": 1.98e-05,
|
| 31 |
+
"loss": 4.7128,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.2,
|
| 36 |
+
"learning_rate": 2.4799999999999996e-05,
|
| 37 |
+
"loss": 4.6691,
|
| 38 |
+
"step": 250
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.24,
|
| 42 |
+
"learning_rate": 2.9799999999999996e-05,
|
| 43 |
+
"loss": 4.6397,
|
| 44 |
+
"step": 300
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.28,
|
| 48 |
+
"learning_rate": 3.48e-05,
|
| 49 |
+
"loss": 4.591,
|
| 50 |
+
"step": 350
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.32,
|
| 54 |
+
"learning_rate": 3.979999999999999e-05,
|
| 55 |
+
"loss": 4.5646,
|
| 56 |
+
"step": 400
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.36,
|
| 60 |
+
"learning_rate": 4.48e-05,
|
| 61 |
+
"loss": 4.5257,
|
| 62 |
+
"step": 450
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.4,
|
| 66 |
+
"learning_rate": 4.98e-05,
|
| 67 |
+
"loss": 4.4901,
|
| 68 |
+
"step": 500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.4,
|
| 72 |
+
"eval_bleu": 0.11633427622704212,
|
| 73 |
+
"eval_loss": 4.419718265533447,
|
| 74 |
+
"eval_runtime": 1914.9377,
|
| 75 |
+
"eval_samples_per_second": 4.6,
|
| 76 |
+
"eval_steps_per_second": 0.144,
|
| 77 |
+
"step": 500
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.44,
|
| 81 |
+
"learning_rate": 5.48e-05,
|
| 82 |
+
"loss": 4.5062,
|
| 83 |
+
"step": 550
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.48,
|
| 87 |
+
"learning_rate": 5.98e-05,
|
| 88 |
+
"loss": 4.4699,
|
| 89 |
+
"step": 600
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.52,
|
| 93 |
+
"learning_rate": 6.479999999999999e-05,
|
| 94 |
+
"loss": 4.4781,
|
| 95 |
+
"step": 650
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.57,
|
| 99 |
+
"learning_rate": 6.979999999999999e-05,
|
| 100 |
+
"loss": 4.433,
|
| 101 |
+
"step": 700
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.61,
|
| 105 |
+
"learning_rate": 7.479999999999999e-05,
|
| 106 |
+
"loss": 4.4437,
|
| 107 |
+
"step": 750
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.65,
|
| 111 |
+
"learning_rate": 7.98e-05,
|
| 112 |
+
"loss": 4.4278,
|
| 113 |
+
"step": 800
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.69,
|
| 117 |
+
"learning_rate": 8.48e-05,
|
| 118 |
+
"loss": 4.4062,
|
| 119 |
+
"step": 850
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.73,
|
| 123 |
+
"learning_rate": 8.98e-05,
|
| 124 |
+
"loss": 4.3717,
|
| 125 |
+
"step": 900
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.77,
|
| 129 |
+
"learning_rate": 9.479999999999999e-05,
|
| 130 |
+
"loss": 4.3586,
|
| 131 |
+
"step": 950
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.81,
|
| 135 |
+
"learning_rate": 9.979999999999999e-05,
|
| 136 |
+
"loss": 4.3929,
|
| 137 |
+
"step": 1000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.81,
|
| 141 |
+
"eval_bleu": 0.15546068077729897,
|
| 142 |
+
"eval_loss": 4.284366607666016,
|
| 143 |
+
"eval_runtime": 1911.4105,
|
| 144 |
+
"eval_samples_per_second": 4.608,
|
| 145 |
+
"eval_steps_per_second": 0.144,
|
| 146 |
+
"step": 1000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.85,
|
| 150 |
+
"learning_rate": 0.00010479999999999999,
|
| 151 |
+
"loss": 4.3491,
|
| 152 |
+
"step": 1050
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.89,
|
| 156 |
+
"learning_rate": 0.00010979999999999999,
|
| 157 |
+
"loss": 4.3284,
|
| 158 |
+
"step": 1100
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.93,
|
| 162 |
+
"learning_rate": 0.00011479999999999999,
|
| 163 |
+
"loss": 4.3209,
|
| 164 |
+
"step": 1150
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.97,
|
| 168 |
+
"learning_rate": 0.00011979999999999998,
|
| 169 |
+
"loss": 4.3172,
|
| 170 |
+
"step": 1200
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.01,
|
| 174 |
+
"learning_rate": 0.00012479999999999997,
|
| 175 |
+
"loss": 4.3263,
|
| 176 |
+
"step": 1250
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.05,
|
| 180 |
+
"learning_rate": 0.00012979999999999998,
|
| 181 |
+
"loss": 4.1922,
|
| 182 |
+
"step": 1300
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.09,
|
| 186 |
+
"learning_rate": 0.00013479999999999997,
|
| 187 |
+
"loss": 4.1936,
|
| 188 |
+
"step": 1350
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.13,
|
| 192 |
+
"learning_rate": 0.00013979999999999998,
|
| 193 |
+
"loss": 4.1718,
|
| 194 |
+
"step": 1400
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.17,
|
| 198 |
+
"learning_rate": 0.0001448,
|
| 199 |
+
"loss": 4.1767,
|
| 200 |
+
"step": 1450
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.21,
|
| 204 |
+
"learning_rate": 0.00014979999999999998,
|
| 205 |
+
"loss": 4.1589,
|
| 206 |
+
"step": 1500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.21,
|
| 210 |
+
"eval_bleu": 0.28584924145284585,
|
| 211 |
+
"eval_loss": 4.1763997077941895,
|
| 212 |
+
"eval_runtime": 1912.3927,
|
| 213 |
+
"eval_samples_per_second": 4.606,
|
| 214 |
+
"eval_steps_per_second": 0.144,
|
| 215 |
+
"step": 1500
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.25,
|
| 219 |
+
"learning_rate": 0.00014999895016087306,
|
| 220 |
+
"loss": 4.1576,
|
| 221 |
+
"step": 1550
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.29,
|
| 225 |
+
"learning_rate": 0.00014999562388001478,
|
| 226 |
+
"loss": 4.1006,
|
| 227 |
+
"step": 1600
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.33,
|
| 231 |
+
"learning_rate": 0.00014999001943582685,
|
| 232 |
+
"loss": 4.1227,
|
| 233 |
+
"step": 1650
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.37,
|
| 237 |
+
"learning_rate": 0.0001499821369985571,
|
| 238 |
+
"loss": 4.1087,
|
| 239 |
+
"step": 1700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.41,
|
| 243 |
+
"learning_rate": 0.0001499719768076526,
|
| 244 |
+
"loss": 4.0728,
|
| 245 |
+
"step": 1750
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.45,
|
| 249 |
+
"learning_rate": 0.00014995953917175227,
|
| 250 |
+
"loss": 4.0705,
|
| 251 |
+
"step": 1800
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.49,
|
| 255 |
+
"learning_rate": 0.00014994482446867774,
|
| 256 |
+
"loss": 4.0425,
|
| 257 |
+
"step": 1850
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.53,
|
| 261 |
+
"learning_rate": 0.00014992783314542174,
|
| 262 |
+
"loss": 4.0156,
|
| 263 |
+
"step": 1900
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.57,
|
| 267 |
+
"learning_rate": 0.00014990856571813448,
|
| 268 |
+
"loss": 4.0317,
|
| 269 |
+
"step": 1950
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.62,
|
| 273 |
+
"learning_rate": 0.0001498870227721081,
|
| 274 |
+
"loss": 3.9819,
|
| 275 |
+
"step": 2000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.62,
|
| 279 |
+
"eval_bleu": 0.39689971475030256,
|
| 280 |
+
"eval_loss": 3.9646406173706055,
|
| 281 |
+
"eval_runtime": 1910.9822,
|
| 282 |
+
"eval_samples_per_second": 4.609,
|
| 283 |
+
"eval_steps_per_second": 0.144,
|
| 284 |
+
"step": 2000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.66,
|
| 288 |
+
"learning_rate": 0.0001498632049617587,
|
| 289 |
+
"loss": 3.9671,
|
| 290 |
+
"step": 2050
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.7,
|
| 294 |
+
"learning_rate": 0.00014983711301060673,
|
| 295 |
+
"loss": 3.9195,
|
| 296 |
+
"step": 2100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.74,
|
| 300 |
+
"learning_rate": 0.00014980874771125478,
|
| 301 |
+
"loss": 3.9144,
|
| 302 |
+
"step": 2150
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.78,
|
| 306 |
+
"learning_rate": 0.00014977810992536359,
|
| 307 |
+
"loss": 3.9092,
|
| 308 |
+
"step": 2200
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.82,
|
| 312 |
+
"learning_rate": 0.00014974520058362584,
|
| 313 |
+
"loss": 3.9027,
|
| 314 |
+
"step": 2250
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.86,
|
| 318 |
+
"learning_rate": 0.00014971002068573793,
|
| 319 |
+
"loss": 3.8732,
|
| 320 |
+
"step": 2300
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.9,
|
| 324 |
+
"learning_rate": 0.00014967257130036961,
|
| 325 |
+
"loss": 3.8439,
|
| 326 |
+
"step": 2350
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.94,
|
| 330 |
+
"learning_rate": 0.00014963285356513152,
|
| 331 |
+
"loss": 3.8375,
|
| 332 |
+
"step": 2400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.98,
|
| 336 |
+
"learning_rate": 0.0001495908686865405,
|
| 337 |
+
"loss": 3.8019,
|
| 338 |
+
"step": 2450
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.02,
|
| 342 |
+
"learning_rate": 0.00014954661793998317,
|
| 343 |
+
"loss": 3.7382,
|
| 344 |
+
"step": 2500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 2.02,
|
| 348 |
+
"eval_bleu": 0.5913530218314237,
|
| 349 |
+
"eval_loss": 3.787797212600708,
|
| 350 |
+
"eval_runtime": 1914.1909,
|
| 351 |
+
"eval_samples_per_second": 4.601,
|
| 352 |
+
"eval_steps_per_second": 0.144,
|
| 353 |
+
"step": 2500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.06,
|
| 357 |
+
"learning_rate": 0.0001495001026696769,
|
| 358 |
+
"loss": 3.5496,
|
| 359 |
+
"step": 2550
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 2.1,
|
| 363 |
+
"learning_rate": 0.00014945132428862936,
|
| 364 |
+
"loss": 3.5282,
|
| 365 |
+
"step": 2600
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 2.14,
|
| 369 |
+
"learning_rate": 0.00014940028427859524,
|
| 370 |
+
"loss": 3.5239,
|
| 371 |
+
"step": 2650
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 2.18,
|
| 375 |
+
"learning_rate": 0.00014934698419003133,
|
| 376 |
+
"loss": 3.4925,
|
| 377 |
+
"step": 2700
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.22,
|
| 381 |
+
"learning_rate": 0.0001492914256420496,
|
| 382 |
+
"loss": 3.5009,
|
| 383 |
+
"step": 2750
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 2.26,
|
| 387 |
+
"learning_rate": 0.00014923361032236776,
|
| 388 |
+
"loss": 3.4963,
|
| 389 |
+
"step": 2800
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 2.3,
|
| 393 |
+
"learning_rate": 0.00014917353998725823,
|
| 394 |
+
"loss": 3.491,
|
| 395 |
+
"step": 2850
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 2.34,
|
| 399 |
+
"learning_rate": 0.00014911121646149456,
|
| 400 |
+
"loss": 3.4631,
|
| 401 |
+
"step": 2900
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 2.38,
|
| 405 |
+
"learning_rate": 0.00014904664163829616,
|
| 406 |
+
"loss": 3.4539,
|
| 407 |
+
"step": 2950
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 2.42,
|
| 411 |
+
"learning_rate": 0.00014897981747927076,
|
| 412 |
+
"loss": 3.4567,
|
| 413 |
+
"step": 3000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 2.42,
|
| 417 |
+
"eval_bleu": 0.7641516646931262,
|
| 418 |
+
"eval_loss": 3.635390520095825,
|
| 419 |
+
"eval_runtime": 1917.6351,
|
| 420 |
+
"eval_samples_per_second": 4.593,
|
| 421 |
+
"eval_steps_per_second": 0.144,
|
| 422 |
+
"step": 3000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.46,
|
| 426 |
+
"learning_rate": 0.00014891074601435482,
|
| 427 |
+
"loss": 3.4458,
|
| 428 |
+
"step": 3050
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.5,
|
| 432 |
+
"learning_rate": 0.00014883942934175177,
|
| 433 |
+
"loss": 3.4317,
|
| 434 |
+
"step": 3100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.54,
|
| 438 |
+
"learning_rate": 0.0001487658696278684,
|
| 439 |
+
"loss": 3.4017,
|
| 440 |
+
"step": 3150
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.58,
|
| 444 |
+
"learning_rate": 0.000148690069107249,
|
| 445 |
+
"loss": 3.4068,
|
| 446 |
+
"step": 3200
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.62,
|
| 450 |
+
"learning_rate": 0.00014861203008250745,
|
| 451 |
+
"loss": 3.3839,
|
| 452 |
+
"step": 3250
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.67,
|
| 456 |
+
"learning_rate": 0.0001485317549242574,
|
| 457 |
+
"loss": 3.4197,
|
| 458 |
+
"step": 3300
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.71,
|
| 462 |
+
"learning_rate": 0.00014844924607104,
|
| 463 |
+
"loss": 3.3836,
|
| 464 |
+
"step": 3350
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.75,
|
| 468 |
+
"learning_rate": 0.00014836450602925014,
|
| 469 |
+
"loss": 3.3611,
|
| 470 |
+
"step": 3400
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.79,
|
| 474 |
+
"learning_rate": 0.00014827753737306008,
|
| 475 |
+
"loss": 3.3541,
|
| 476 |
+
"step": 3450
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.83,
|
| 480 |
+
"learning_rate": 0.00014818834274434134,
|
| 481 |
+
"loss": 3.3587,
|
| 482 |
+
"step": 3500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.83,
|
| 486 |
+
"eval_bleu": 1.1491131179238592,
|
| 487 |
+
"eval_loss": 3.5020267963409424,
|
| 488 |
+
"eval_runtime": 1914.8223,
|
| 489 |
+
"eval_samples_per_second": 4.6,
|
| 490 |
+
"eval_steps_per_second": 0.144,
|
| 491 |
+
"step": 3500
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 2.87,
|
| 495 |
+
"learning_rate": 0.00014809692485258445,
|
| 496 |
+
"loss": 3.3124,
|
| 497 |
+
"step": 3550
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 2.91,
|
| 501 |
+
"learning_rate": 0.00014800328647481662,
|
| 502 |
+
"loss": 3.3311,
|
| 503 |
+
"step": 3600
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 2.95,
|
| 507 |
+
"learning_rate": 0.00014790743045551744,
|
| 508 |
+
"loss": 3.302,
|
| 509 |
+
"step": 3650
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 2.99,
|
| 513 |
+
"learning_rate": 0.00014780935970653235,
|
| 514 |
+
"loss": 3.2845,
|
| 515 |
+
"step": 3700
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.03,
|
| 519 |
+
"learning_rate": 0.00014770907720698426,
|
| 520 |
+
"loss": 3.165,
|
| 521 |
+
"step": 3750
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.07,
|
| 525 |
+
"learning_rate": 0.00014760658600318318,
|
| 526 |
+
"loss": 3.0311,
|
| 527 |
+
"step": 3800
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.11,
|
| 531 |
+
"learning_rate": 0.00014750188920853338,
|
| 532 |
+
"loss": 3.0383,
|
| 533 |
+
"step": 3850
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 3.15,
|
| 537 |
+
"learning_rate": 0.00014739499000343914,
|
| 538 |
+
"loss": 3.0399,
|
| 539 |
+
"step": 3900
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.19,
|
| 543 |
+
"learning_rate": 0.0001472858916352079,
|
| 544 |
+
"loss": 3.0327,
|
| 545 |
+
"step": 3950
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.23,
|
| 549 |
+
"learning_rate": 0.0001471745974179517,
|
| 550 |
+
"loss": 3.007,
|
| 551 |
+
"step": 4000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.23,
|
| 555 |
+
"eval_bleu": 1.106589199275075,
|
| 556 |
+
"eval_loss": 3.435346841812134,
|
| 557 |
+
"eval_runtime": 1918.0445,
|
| 558 |
+
"eval_samples_per_second": 4.592,
|
| 559 |
+
"eval_steps_per_second": 0.144,
|
| 560 |
+
"step": 4000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 3.27,
|
| 564 |
+
"learning_rate": 0.00014706111073248656,
|
| 565 |
+
"loss": 3.0285,
|
| 566 |
+
"step": 4050
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.31,
|
| 570 |
+
"learning_rate": 0.0001469454350262297,
|
| 571 |
+
"loss": 3.0335,
|
| 572 |
+
"step": 4100
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.35,
|
| 576 |
+
"learning_rate": 0.0001468275738130948,
|
| 577 |
+
"loss": 3.0331,
|
| 578 |
+
"step": 4150
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.39,
|
| 582 |
+
"learning_rate": 0.0001467075306733854,
|
| 583 |
+
"loss": 3.024,
|
| 584 |
+
"step": 4200
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.43,
|
| 588 |
+
"learning_rate": 0.000146585309253686,
|
| 589 |
+
"loss": 3.0199,
|
| 590 |
+
"step": 4250
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.47,
|
| 594 |
+
"learning_rate": 0.00014646091326675126,
|
| 595 |
+
"loss": 2.9867,
|
| 596 |
+
"step": 4300
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.51,
|
| 600 |
+
"learning_rate": 0.00014633434649139344,
|
| 601 |
+
"loss": 3.0297,
|
| 602 |
+
"step": 4350
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.55,
|
| 606 |
+
"learning_rate": 0.00014620561277236722,
|
| 607 |
+
"loss": 2.9971,
|
| 608 |
+
"step": 4400
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.59,
|
| 612 |
+
"learning_rate": 0.0001460747160202534,
|
| 613 |
+
"loss": 2.9872,
|
| 614 |
+
"step": 4450
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.63,
|
| 618 |
+
"learning_rate": 0.0001459416602113397,
|
| 619 |
+
"loss": 2.9901,
|
| 620 |
+
"step": 4500
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.63,
|
| 624 |
+
"eval_bleu": 1.4837165956404195,
|
| 625 |
+
"eval_loss": 3.3302574157714844,
|
| 626 |
+
"eval_runtime": 1928.4326,
|
| 627 |
+
"eval_samples_per_second": 4.567,
|
| 628 |
+
"eval_steps_per_second": 0.143,
|
| 629 |
+
"step": 4500
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.67,
|
| 633 |
+
"learning_rate": 0.00014580644938750012,
|
| 634 |
+
"loss": 2.9842,
|
| 635 |
+
"step": 4550
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 3.72,
|
| 639 |
+
"learning_rate": 0.00014566908765607222,
|
| 640 |
+
"loss": 2.9818,
|
| 641 |
+
"step": 4600
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 3.76,
|
| 645 |
+
"learning_rate": 0.00014552957918973226,
|
| 646 |
+
"loss": 2.9784,
|
| 647 |
+
"step": 4650
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 3.8,
|
| 651 |
+
"learning_rate": 0.00014538792822636849,
|
| 652 |
+
"loss": 2.974,
|
| 653 |
+
"step": 4700
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 3.84,
|
| 657 |
+
"learning_rate": 0.00014524413906895234,
|
| 658 |
+
"loss": 2.9592,
|
| 659 |
+
"step": 4750
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 3.88,
|
| 663 |
+
"learning_rate": 0.00014509821608540784,
|
| 664 |
+
"loss": 2.9519,
|
| 665 |
+
"step": 4800
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 3.92,
|
| 669 |
+
"learning_rate": 0.00014495016370847882,
|
| 670 |
+
"loss": 2.9672,
|
| 671 |
+
"step": 4850
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 3.96,
|
| 675 |
+
"learning_rate": 0.00014479998643559435,
|
| 676 |
+
"loss": 2.944,
|
| 677 |
+
"step": 4900
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 4.0,
|
| 681 |
+
"learning_rate": 0.00014464768882873198,
|
| 682 |
+
"loss": 2.9653,
|
| 683 |
+
"step": 4950
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 4.04,
|
| 687 |
+
"learning_rate": 0.00014449327551427935,
|
| 688 |
+
"loss": 2.7312,
|
| 689 |
+
"step": 5000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 4.04,
|
| 693 |
+
"eval_bleu": 1.3796115530850706,
|
| 694 |
+
"eval_loss": 3.2999939918518066,
|
| 695 |
+
"eval_runtime": 1915.5145,
|
| 696 |
+
"eval_samples_per_second": 4.598,
|
| 697 |
+
"eval_steps_per_second": 0.144,
|
| 698 |
+
"step": 5000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.08,
|
| 702 |
+
"learning_rate": 0.0001443367511828934,
|
| 703 |
+
"loss": 2.6928,
|
| 704 |
+
"step": 5050
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 4.12,
|
| 708 |
+
"learning_rate": 0.0001441781205893582,
|
| 709 |
+
"loss": 2.6899,
|
| 710 |
+
"step": 5100
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 4.16,
|
| 714 |
+
"learning_rate": 0.00014401738855244028,
|
| 715 |
+
"loss": 2.6697,
|
| 716 |
+
"step": 5150
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.2,
|
| 720 |
+
"learning_rate": 0.00014385455995474222,
|
| 721 |
+
"loss": 2.6839,
|
| 722 |
+
"step": 5200
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.24,
|
| 726 |
+
"learning_rate": 0.00014368963974255454,
|
| 727 |
+
"loss": 2.6911,
|
| 728 |
+
"step": 5250
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.28,
|
| 732 |
+
"learning_rate": 0.0001435226329257053,
|
| 733 |
+
"loss": 2.7043,
|
| 734 |
+
"step": 5300
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.32,
|
| 738 |
+
"learning_rate": 0.00014335354457740792,
|
| 739 |
+
"loss": 2.6771,
|
| 740 |
+
"step": 5350
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.36,
|
| 744 |
+
"learning_rate": 0.00014318237983410706,
|
| 745 |
+
"loss": 2.6987,
|
| 746 |
+
"step": 5400
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.4,
|
| 750 |
+
"learning_rate": 0.0001430091438953227,
|
| 751 |
+
"loss": 2.6854,
|
| 752 |
+
"step": 5450
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.44,
|
| 756 |
+
"learning_rate": 0.00014283384202349203,
|
| 757 |
+
"loss": 2.6822,
|
| 758 |
+
"step": 5500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.44,
|
| 762 |
+
"eval_bleu": 1.572420203309693,
|
| 763 |
+
"eval_loss": 3.2495412826538086,
|
| 764 |
+
"eval_runtime": 1922.8389,
|
| 765 |
+
"eval_samples_per_second": 4.581,
|
| 766 |
+
"eval_steps_per_second": 0.144,
|
| 767 |
+
"step": 5500
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 4.48,
|
| 771 |
+
"learning_rate": 0.00014265647954380976,
|
| 772 |
+
"loss": 2.7064,
|
| 773 |
+
"step": 5550
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 4.52,
|
| 777 |
+
"learning_rate": 0.00014247706184406618,
|
| 778 |
+
"loss": 2.6905,
|
| 779 |
+
"step": 5600
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 4.56,
|
| 783 |
+
"learning_rate": 0.00014229559437448362,
|
| 784 |
+
"loss": 2.6847,
|
| 785 |
+
"step": 5650
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 4.6,
|
| 789 |
+
"learning_rate": 0.00014211208264755092,
|
| 790 |
+
"loss": 2.6825,
|
| 791 |
+
"step": 5700
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 4.64,
|
| 795 |
+
"learning_rate": 0.00014192653223785577,
|
| 796 |
+
"loss": 2.6759,
|
| 797 |
+
"step": 5750
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 4.68,
|
| 801 |
+
"learning_rate": 0.0001417389487819156,
|
| 802 |
+
"loss": 2.6888,
|
| 803 |
+
"step": 5800
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 4.72,
|
| 807 |
+
"learning_rate": 0.00014154933797800621,
|
| 808 |
+
"loss": 2.6722,
|
| 809 |
+
"step": 5850
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 4.77,
|
| 813 |
+
"learning_rate": 0.0001413577055859888,
|
| 814 |
+
"loss": 2.688,
|
| 815 |
+
"step": 5900
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 4.81,
|
| 819 |
+
"learning_rate": 0.00014116405742713484,
|
| 820 |
+
"loss": 2.6934,
|
| 821 |
+
"step": 5950
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 4.85,
|
| 825 |
+
"learning_rate": 0.00014096839938394936,
|
| 826 |
+
"loss": 2.6577,
|
| 827 |
+
"step": 6000
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 4.85,
|
| 831 |
+
"eval_bleu": 1.6301048703044616,
|
| 832 |
+
"eval_loss": 3.176309823989868,
|
| 833 |
+
"eval_runtime": 1919.9365,
|
| 834 |
+
"eval_samples_per_second": 4.588,
|
| 835 |
+
"eval_steps_per_second": 0.144,
|
| 836 |
+
"step": 6000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.89,
|
| 840 |
+
"learning_rate": 0.00014077073739999222,
|
| 841 |
+
"loss": 2.6749,
|
| 842 |
+
"step": 6050
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.93,
|
| 846 |
+
"learning_rate": 0.0001405710774796975,
|
| 847 |
+
"loss": 2.6639,
|
| 848 |
+
"step": 6100
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.97,
|
| 852 |
+
"learning_rate": 0.0001403694256881913,
|
| 853 |
+
"loss": 2.6828,
|
| 854 |
+
"step": 6150
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 5.01,
|
| 858 |
+
"learning_rate": 0.00014016578815110716,
|
| 859 |
+
"loss": 2.645,
|
| 860 |
+
"step": 6200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 5.05,
|
| 864 |
+
"learning_rate": 0.00013996017105440036,
|
| 865 |
+
"loss": 2.3837,
|
| 866 |
+
"step": 6250
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 5.09,
|
| 870 |
+
"learning_rate": 0.00013975258064415972,
|
| 871 |
+
"loss": 2.3875,
|
| 872 |
+
"step": 6300
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 5.13,
|
| 876 |
+
"learning_rate": 0.00013954302322641797,
|
| 877 |
+
"loss": 2.3975,
|
| 878 |
+
"step": 6350
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 5.17,
|
| 882 |
+
"learning_rate": 0.00013933150516696024,
|
| 883 |
+
"loss": 2.4332,
|
| 884 |
+
"step": 6400
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 5.21,
|
| 888 |
+
"learning_rate": 0.00013911803289113055,
|
| 889 |
+
"loss": 2.3976,
|
| 890 |
+
"step": 6450
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 5.25,
|
| 894 |
+
"learning_rate": 0.00013890261288363676,
|
| 895 |
+
"loss": 2.4118,
|
| 896 |
+
"step": 6500
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 5.25,
|
| 900 |
+
"eval_bleu": 1.5739999205693251,
|
| 901 |
+
"eval_loss": 3.2138876914978027,
|
| 902 |
+
"eval_runtime": 1926.1281,
|
| 903 |
+
"eval_samples_per_second": 4.573,
|
| 904 |
+
"eval_steps_per_second": 0.143,
|
| 905 |
+
"step": 6500
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 5.29,
|
| 909 |
+
"learning_rate": 0.00013868525168835353,
|
| 910 |
+
"loss": 2.4436,
|
| 911 |
+
"step": 6550
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 5.33,
|
| 915 |
+
"learning_rate": 0.0001384659559081235,
|
| 916 |
+
"loss": 2.428,
|
| 917 |
+
"step": 6600
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 5.37,
|
| 921 |
+
"learning_rate": 0.0001382447322045568,
|
| 922 |
+
"loss": 2.4192,
|
| 923 |
+
"step": 6650
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 5.41,
|
| 927 |
+
"learning_rate": 0.0001380215872978285,
|
| 928 |
+
"loss": 2.4337,
|
| 929 |
+
"step": 6700
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 5.45,
|
| 933 |
+
"learning_rate": 0.0001377965279664748,
|
| 934 |
+
"loss": 2.4193,
|
| 935 |
+
"step": 6750
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 5.49,
|
| 939 |
+
"learning_rate": 0.0001375695610471868,
|
| 940 |
+
"loss": 2.438,
|
| 941 |
+
"step": 6800
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 5.53,
|
| 945 |
+
"learning_rate": 0.00013734069343460293,
|
| 946 |
+
"loss": 2.4385,
|
| 947 |
+
"step": 6850
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 5.57,
|
| 951 |
+
"learning_rate": 0.0001371099320810995,
|
| 952 |
+
"loss": 2.4556,
|
| 953 |
+
"step": 6900
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 5.61,
|
| 957 |
+
"learning_rate": 0.0001368772839965797,
|
| 958 |
+
"loss": 2.4518,
|
| 959 |
+
"step": 6950
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 5.65,
|
| 963 |
+
"learning_rate": 0.00013664275624826025,
|
| 964 |
+
"loss": 2.4345,
|
| 965 |
+
"step": 7000
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 5.65,
|
| 969 |
+
"eval_bleu": 1.9002943282055358,
|
| 970 |
+
"eval_loss": 3.1631877422332764,
|
| 971 |
+
"eval_runtime": 1923.9555,
|
| 972 |
+
"eval_samples_per_second": 4.578,
|
| 973 |
+
"eval_steps_per_second": 0.143,
|
| 974 |
+
"step": 7000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.69,
|
| 978 |
+
"learning_rate": 0.00013640635596045707,
|
| 979 |
+
"loss": 2.4438,
|
| 980 |
+
"step": 7050
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.73,
|
| 984 |
+
"learning_rate": 0.00013616809031436876,
|
| 985 |
+
"loss": 2.4239,
|
| 986 |
+
"step": 7100
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.77,
|
| 990 |
+
"learning_rate": 0.0001359279665478584,
|
| 991 |
+
"loss": 2.4293,
|
| 992 |
+
"step": 7150
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.82,
|
| 996 |
+
"learning_rate": 0.0001356859919552337,
|
| 997 |
+
"loss": 2.4569,
|
| 998 |
+
"step": 7200
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.86,
|
| 1002 |
+
"learning_rate": 0.0001354421738870255,
|
| 1003 |
+
"loss": 2.4437,
|
| 1004 |
+
"step": 7250
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.9,
|
| 1008 |
+
"learning_rate": 0.00013519651974976433,
|
| 1009 |
+
"loss": 2.4373,
|
| 1010 |
+
"step": 7300
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.94,
|
| 1014 |
+
"learning_rate": 0.00013494903700575562,
|
| 1015 |
+
"loss": 2.451,
|
| 1016 |
+
"step": 7350
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.98,
|
| 1020 |
+
"learning_rate": 0.00013469973317285284,
|
| 1021 |
+
"loss": 2.4468,
|
| 1022 |
+
"step": 7400
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 6.02,
|
| 1026 |
+
"learning_rate": 0.0001344486158242292,
|
| 1027 |
+
"loss": 2.333,
|
| 1028 |
+
"step": 7450
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 6.06,
|
| 1032 |
+
"learning_rate": 0.00013419569258814757,
|
| 1033 |
+
"loss": 2.158,
|
| 1034 |
+
"step": 7500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 6.06,
|
| 1038 |
+
"eval_bleu": 1.8111395024627068,
|
| 1039 |
+
"eval_loss": 3.179304361343384,
|
| 1040 |
+
"eval_runtime": 1921.8485,
|
| 1041 |
+
"eval_samples_per_second": 4.583,
|
| 1042 |
+
"eval_steps_per_second": 0.144,
|
| 1043 |
+
"step": 7500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 6.1,
|
| 1047 |
+
"learning_rate": 0.00013394097114772887,
|
| 1048 |
+
"loss": 2.1613,
|
| 1049 |
+
"step": 7550
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 6.14,
|
| 1053 |
+
"learning_rate": 0.00013368445924071844,
|
| 1054 |
+
"loss": 2.1807,
|
| 1055 |
+
"step": 7600
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 6.18,
|
| 1059 |
+
"learning_rate": 0.00013342616465925126,
|
| 1060 |
+
"loss": 2.1873,
|
| 1061 |
+
"step": 7650
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 6.22,
|
| 1065 |
+
"learning_rate": 0.00013316609524961502,
|
| 1066 |
+
"loss": 2.1955,
|
| 1067 |
+
"step": 7700
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 6.26,
|
| 1071 |
+
"learning_rate": 0.00013290425891201196,
|
| 1072 |
+
"loss": 2.2115,
|
| 1073 |
+
"step": 7750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 6.3,
|
| 1077 |
+
"learning_rate": 0.00013264066360031872,
|
| 1078 |
+
"loss": 2.1964,
|
| 1079 |
+
"step": 7800
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 6.34,
|
| 1083 |
+
"learning_rate": 0.0001323753173218448,
|
| 1084 |
+
"loss": 2.2045,
|
| 1085 |
+
"step": 7850
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 6.38,
|
| 1089 |
+
"learning_rate": 0.00013210822813708936,
|
| 1090 |
+
"loss": 2.1907,
|
| 1091 |
+
"step": 7900
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 6.42,
|
| 1095 |
+
"learning_rate": 0.0001318394041594963,
|
| 1096 |
+
"loss": 2.2066,
|
| 1097 |
+
"step": 7950
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 6.46,
|
| 1101 |
+
"learning_rate": 0.00013156885355520778,
|
| 1102 |
+
"loss": 2.2051,
|
| 1103 |
+
"step": 8000
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 6.46,
|
| 1107 |
+
"eval_bleu": 1.98361849433367,
|
| 1108 |
+
"eval_loss": 3.160503625869751,
|
| 1109 |
+
"eval_runtime": 1922.8791,
|
| 1110 |
+
"eval_samples_per_second": 4.581,
|
| 1111 |
+
"eval_steps_per_second": 0.144,
|
| 1112 |
+
"step": 8000
|
| 1113 |
+
}
|
| 1114 |
+
],
|
| 1115 |
+
"max_steps": 30000,
|
| 1116 |
+
"num_train_epochs": 25,
|
| 1117 |
+
"total_flos": 4.9816432750591004e+20,
|
| 1118 |
+
"trial_name": null,
|
| 1119 |
+
"trial_params": null
|
| 1120 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d60540c0ce430e5b6261166a4050827eb5d7605940364f01a54b5855933773f5
|
| 3 |
+
size 4024
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|