Upload folder using huggingface_hub
Browse files- checkpoint-18018/config.json +71 -0
- checkpoint-18018/generation_config.json +7 -0
- checkpoint-18018/model.safetensors +3 -0
- checkpoint-18018/optimizer.pt +3 -0
- checkpoint-18018/rng_state.pth +3 -0
- checkpoint-18018/scheduler.pt +3 -0
- checkpoint-18018/trainer_state.json +320 -0
- checkpoint-18018/training_args.bin +3 -0
- config.json +71 -0
- events.out.tfevents.1751355165.gra5 +3 -0
- generation_config.json +7 -0
- model.safetensors +3 -0
- results.json +17 -0
- training_args.bin +3 -0
checkpoint-18018/config.json
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/home/mootez/projects/def-tusharma/mootez/local_models/codet5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"classifier_dropout": 0.0,
|
| 8 |
+
"d_ff": 3072,
|
| 9 |
+
"d_kv": 64,
|
| 10 |
+
"d_model": 768,
|
| 11 |
+
"decoder_start_token_id": 0,
|
| 12 |
+
"dense_act_fn": "relu",
|
| 13 |
+
"dropout_rate": 0.1,
|
| 14 |
+
"eos_token_id": 2,
|
| 15 |
+
"feed_forward_proj": "relu",
|
| 16 |
+
"gradient_checkpointing": false,
|
| 17 |
+
"id2label": {
|
| 18 |
+
"0": "LABEL_0"
|
| 19 |
+
},
|
| 20 |
+
"initializer_factor": 1.0,
|
| 21 |
+
"is_encoder_decoder": true,
|
| 22 |
+
"is_gated_act": false,
|
| 23 |
+
"label2id": {
|
| 24 |
+
"LABEL_0": 0
|
| 25 |
+
},
|
| 26 |
+
"layer_norm_epsilon": 1e-06,
|
| 27 |
+
"merge_layer_index": 5,
|
| 28 |
+
"merger_type": "mean",
|
| 29 |
+
"model_type": "t5",
|
| 30 |
+
"n_positions": 512,
|
| 31 |
+
"num_decoder_layers": 12,
|
| 32 |
+
"num_heads": 12,
|
| 33 |
+
"num_layers": 12,
|
| 34 |
+
"output_past": true,
|
| 35 |
+
"pad_token_id": 0,
|
| 36 |
+
"relative_attention_max_distance": 128,
|
| 37 |
+
"relative_attention_num_buckets": 32,
|
| 38 |
+
"task_specific_params": {
|
| 39 |
+
"summarization": {
|
| 40 |
+
"early_stopping": true,
|
| 41 |
+
"length_penalty": 2.0,
|
| 42 |
+
"max_length": 200,
|
| 43 |
+
"min_length": 30,
|
| 44 |
+
"no_repeat_ngram_size": 3,
|
| 45 |
+
"num_beams": 4,
|
| 46 |
+
"prefix": "summarize: "
|
| 47 |
+
},
|
| 48 |
+
"translation_en_to_de": {
|
| 49 |
+
"early_stopping": true,
|
| 50 |
+
"max_length": 300,
|
| 51 |
+
"num_beams": 4,
|
| 52 |
+
"prefix": "translate English to German: "
|
| 53 |
+
},
|
| 54 |
+
"translation_en_to_fr": {
|
| 55 |
+
"early_stopping": true,
|
| 56 |
+
"max_length": 300,
|
| 57 |
+
"num_beams": 4,
|
| 58 |
+
"prefix": "translate English to French: "
|
| 59 |
+
},
|
| 60 |
+
"translation_en_to_ro": {
|
| 61 |
+
"early_stopping": true,
|
| 62 |
+
"max_length": 300,
|
| 63 |
+
"num_beams": 4,
|
| 64 |
+
"prefix": "translate English to Romanian: "
|
| 65 |
+
}
|
| 66 |
+
},
|
| 67 |
+
"torch_dtype": "float32",
|
| 68 |
+
"transformers_version": "4.48.2",
|
| 69 |
+
"use_cache": true,
|
| 70 |
+
"vocab_size": 32100
|
| 71 |
+
}
|
checkpoint-18018/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 1,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.2"
|
| 7 |
+
}
|
checkpoint-18018/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bcf62040e13cb3e83c0bc72a75615abe1dbcf8cc9bcaca08ec263e0552ee355a
|
| 3 |
+
size 891558696
|
checkpoint-18018/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6dffd124588cf932e15f9b63f0d86008487f219cd0f710e10ea833240082232a
|
| 3 |
+
size 1783272762
|
checkpoint-18018/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f22685425876a093316c994a4f9935ce3cfa729e203b1cce6ddd551c2dbd9e0c
|
| 3 |
+
size 14372
|
checkpoint-18018/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e9e2286ca963245683762c024985c3ce1e71e1a1837df88ec9d63c418a60967
|
| 3 |
+
size 1064
|
checkpoint-18018/trainer_state.json
ADDED
|
@@ -0,0 +1,320 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.18638537802601654,
|
| 3 |
+
"best_model_checkpoint": "./logs/mean_codet5-base_L5_codetrans_123456_src_java_tgt_c_sharp/checkpoint-18018",
|
| 4 |
+
"epoch": 7.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 18018,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.3885003885003885,
|
| 13 |
+
"grad_norm": 4.583204746246338,
|
| 14 |
+
"learning_rate": 7.77302759424796e-06,
|
| 15 |
+
"loss": 2.9281,
|
| 16 |
+
"step": 1000
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.777000777000777,
|
| 20 |
+
"grad_norm": 3.8672478199005127,
|
| 21 |
+
"learning_rate": 1.554605518849592e-05,
|
| 22 |
+
"loss": 0.6014,
|
| 23 |
+
"step": 2000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.0,
|
| 27 |
+
"eval_bleu": 0.17871495324307252,
|
| 28 |
+
"eval_brevity_penalty": 0.20143686676996223,
|
| 29 |
+
"eval_length_ratio": 0.38427851085426146,
|
| 30 |
+
"eval_loss": 0.3183501362800598,
|
| 31 |
+
"eval_precision": [
|
| 32 |
+
0.9376504420066035,
|
| 33 |
+
0.8950609309736037,
|
| 34 |
+
0.8690012405597753,
|
| 35 |
+
0.8495203309276536
|
| 36 |
+
],
|
| 37 |
+
"eval_reference_length": 122164,
|
| 38 |
+
"eval_rouge1": 0.560008325932495,
|
| 39 |
+
"eval_rouge2": 0.4767949972409239,
|
| 40 |
+
"eval_rougeL": 0.5578633065205703,
|
| 41 |
+
"eval_rougeLsum": 0.5578849196253808,
|
| 42 |
+
"eval_runtime": 18.9873,
|
| 43 |
+
"eval_samples_per_second": 26.281,
|
| 44 |
+
"eval_steps_per_second": 1.685,
|
| 45 |
+
"eval_translation_length": 46945,
|
| 46 |
+
"step": 2574
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 1.1655011655011656,
|
| 50 |
+
"grad_norm": 4.169002532958984,
|
| 51 |
+
"learning_rate": 1.9631372210471794e-05,
|
| 52 |
+
"loss": 0.4056,
|
| 53 |
+
"step": 3000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.554001554001554,
|
| 57 |
+
"grad_norm": 1.9244877099990845,
|
| 58 |
+
"learning_rate": 1.8768075279492384e-05,
|
| 59 |
+
"loss": 0.3153,
|
| 60 |
+
"step": 4000
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.9425019425019425,
|
| 64 |
+
"grad_norm": 2.8535172939300537,
|
| 65 |
+
"learning_rate": 1.790477834851297e-05,
|
| 66 |
+
"loss": 0.2896,
|
| 67 |
+
"step": 5000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 2.0,
|
| 71 |
+
"eval_bleu": 0.18455277242649967,
|
| 72 |
+
"eval_brevity_penalty": 0.20278901555719767,
|
| 73 |
+
"eval_length_ratio": 0.38526898267902165,
|
| 74 |
+
"eval_loss": 0.227849081158638,
|
| 75 |
+
"eval_precision": [
|
| 76 |
+
0.9466493859686398,
|
| 77 |
+
0.9156269461206433,
|
| 78 |
+
0.8966527741599375,
|
| 79 |
+
0.8826175689613553
|
| 80 |
+
],
|
| 81 |
+
"eval_reference_length": 122164,
|
| 82 |
+
"eval_rouge1": 0.5787394137026639,
|
| 83 |
+
"eval_rouge2": 0.5095173532086278,
|
| 84 |
+
"eval_rougeL": 0.5774934168556525,
|
| 85 |
+
"eval_rougeLsum": 0.5774734498473086,
|
| 86 |
+
"eval_runtime": 18.854,
|
| 87 |
+
"eval_samples_per_second": 26.466,
|
| 88 |
+
"eval_steps_per_second": 1.697,
|
| 89 |
+
"eval_translation_length": 47066,
|
| 90 |
+
"step": 5148
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 2.331002331002331,
|
| 94 |
+
"grad_norm": 4.124932289123535,
|
| 95 |
+
"learning_rate": 1.7041481417533562e-05,
|
| 96 |
+
"loss": 0.2233,
|
| 97 |
+
"step": 6000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 2.7195027195027195,
|
| 101 |
+
"grad_norm": 2.769991397857666,
|
| 102 |
+
"learning_rate": 1.6178184486554153e-05,
|
| 103 |
+
"loss": 0.2138,
|
| 104 |
+
"step": 7000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 3.0,
|
| 108 |
+
"eval_bleu": 0.18207050268413058,
|
| 109 |
+
"eval_brevity_penalty": 0.19739310737467283,
|
| 110 |
+
"eval_length_ratio": 0.38130709537998103,
|
| 111 |
+
"eval_loss": 0.1899239420890808,
|
| 112 |
+
"eval_precision": [
|
| 113 |
+
0.9519986260787429,
|
| 114 |
+
0.9269144803940715,
|
| 115 |
+
0.9113724113724113,
|
| 116 |
+
0.9000332704890762
|
| 117 |
+
],
|
| 118 |
+
"eval_reference_length": 122164,
|
| 119 |
+
"eval_rouge1": 0.5897108321092976,
|
| 120 |
+
"eval_rouge2": 0.5257239528586701,
|
| 121 |
+
"eval_rougeL": 0.5872456030956406,
|
| 122 |
+
"eval_rougeLsum": 0.587488305635381,
|
| 123 |
+
"eval_runtime": 18.975,
|
| 124 |
+
"eval_samples_per_second": 26.298,
|
| 125 |
+
"eval_steps_per_second": 1.686,
|
| 126 |
+
"eval_translation_length": 46582,
|
| 127 |
+
"step": 7722
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.108003108003108,
|
| 131 |
+
"grad_norm": 0.9152127504348755,
|
| 132 |
+
"learning_rate": 1.5314887555574744e-05,
|
| 133 |
+
"loss": 0.1926,
|
| 134 |
+
"step": 8000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.4965034965034967,
|
| 138 |
+
"grad_norm": 0.940152645111084,
|
| 139 |
+
"learning_rate": 1.4451590624595331e-05,
|
| 140 |
+
"loss": 0.1551,
|
| 141 |
+
"step": 9000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.885003885003885,
|
| 145 |
+
"grad_norm": 2.528108835220337,
|
| 146 |
+
"learning_rate": 1.358829369361592e-05,
|
| 147 |
+
"loss": 0.1576,
|
| 148 |
+
"step": 10000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 4.0,
|
| 152 |
+
"eval_bleu": 0.18448253018202196,
|
| 153 |
+
"eval_brevity_penalty": 0.19928448489108977,
|
| 154 |
+
"eval_length_ratio": 0.3826986673651812,
|
| 155 |
+
"eval_loss": 0.1749277263879776,
|
| 156 |
+
"eval_precision": [
|
| 157 |
+
0.952579568788501,
|
| 158 |
+
0.9299504896979656,
|
| 159 |
+
0.9155046553306815,
|
| 160 |
+
0.9055352999668544
|
| 161 |
+
],
|
| 162 |
+
"eval_reference_length": 122164,
|
| 163 |
+
"eval_rouge1": 0.5927507535776158,
|
| 164 |
+
"eval_rouge2": 0.5329461327075656,
|
| 165 |
+
"eval_rougeL": 0.5910730578791388,
|
| 166 |
+
"eval_rougeLsum": 0.5899509488601137,
|
| 167 |
+
"eval_runtime": 18.9307,
|
| 168 |
+
"eval_samples_per_second": 26.359,
|
| 169 |
+
"eval_steps_per_second": 1.69,
|
| 170 |
+
"eval_translation_length": 46752,
|
| 171 |
+
"step": 10296
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 4.273504273504273,
|
| 175 |
+
"grad_norm": 0.4865892231464386,
|
| 176 |
+
"learning_rate": 1.272499676263651e-05,
|
| 177 |
+
"loss": 0.137,
|
| 178 |
+
"step": 11000
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 4.662004662004662,
|
| 182 |
+
"grad_norm": 1.2012805938720703,
|
| 183 |
+
"learning_rate": 1.1861699831657098e-05,
|
| 184 |
+
"loss": 0.1292,
|
| 185 |
+
"step": 12000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 5.0,
|
| 189 |
+
"eval_bleu": 0.1847412745442132,
|
| 190 |
+
"eval_brevity_penalty": 0.19792671160459022,
|
| 191 |
+
"eval_length_ratio": 0.38170000982286106,
|
| 192 |
+
"eval_loss": 0.17231304943561554,
|
| 193 |
+
"eval_precision": [
|
| 194 |
+
0.9574522839373794,
|
| 195 |
+
0.9369621295874792,
|
| 196 |
+
0.9242636746143057,
|
| 197 |
+
0.9153834223295593
|
| 198 |
+
],
|
| 199 |
+
"eval_reference_length": 122164,
|
| 200 |
+
"eval_rouge1": 0.5999162849722973,
|
| 201 |
+
"eval_rouge2": 0.5425490931927184,
|
| 202 |
+
"eval_rougeL": 0.5979426936404479,
|
| 203 |
+
"eval_rougeLsum": 0.5979704973270867,
|
| 204 |
+
"eval_runtime": 18.9618,
|
| 205 |
+
"eval_samples_per_second": 26.316,
|
| 206 |
+
"eval_steps_per_second": 1.688,
|
| 207 |
+
"eval_translation_length": 46630,
|
| 208 |
+
"step": 12870
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 5.05050505050505,
|
| 212 |
+
"grad_norm": 0.9589797258377075,
|
| 213 |
+
"learning_rate": 1.0998402900677689e-05,
|
| 214 |
+
"loss": 0.1234,
|
| 215 |
+
"step": 13000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 5.439005439005439,
|
| 219 |
+
"grad_norm": 0.3940868675708771,
|
| 220 |
+
"learning_rate": 1.0135105969698278e-05,
|
| 221 |
+
"loss": 0.1056,
|
| 222 |
+
"step": 14000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 5.827505827505828,
|
| 226 |
+
"grad_norm": 0.9719197750091553,
|
| 227 |
+
"learning_rate": 9.271809038718868e-06,
|
| 228 |
+
"loss": 0.1072,
|
| 229 |
+
"step": 15000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 6.0,
|
| 233 |
+
"eval_bleu": 0.18501585656019542,
|
| 234 |
+
"eval_brevity_penalty": 0.19827151211459745,
|
| 235 |
+
"eval_length_ratio": 0.3819537670672211,
|
| 236 |
+
"eval_loss": 0.1678522527217865,
|
| 237 |
+
"eval_precision": [
|
| 238 |
+
0.9555303143953194,
|
| 239 |
+
0.936332914518435,
|
| 240 |
+
0.9245997853842279,
|
| 241 |
+
0.9165707200425117
|
| 242 |
+
],
|
| 243 |
+
"eval_reference_length": 122164,
|
| 244 |
+
"eval_rouge1": 0.5995115665557844,
|
| 245 |
+
"eval_rouge2": 0.5438181042938635,
|
| 246 |
+
"eval_rougeL": 0.5980098530810114,
|
| 247 |
+
"eval_rougeLsum": 0.5980445546648558,
|
| 248 |
+
"eval_runtime": 18.8802,
|
| 249 |
+
"eval_samples_per_second": 26.43,
|
| 250 |
+
"eval_steps_per_second": 1.695,
|
| 251 |
+
"eval_translation_length": 46661,
|
| 252 |
+
"step": 15444
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 6.216006216006216,
|
| 256 |
+
"grad_norm": 0.9491456151008606,
|
| 257 |
+
"learning_rate": 8.408512107739457e-06,
|
| 258 |
+
"loss": 0.0946,
|
| 259 |
+
"step": 16000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 6.6045066045066045,
|
| 263 |
+
"grad_norm": 1.224445104598999,
|
| 264 |
+
"learning_rate": 7.545215176760047e-06,
|
| 265 |
+
"loss": 0.0905,
|
| 266 |
+
"step": 17000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 6.993006993006993,
|
| 270 |
+
"grad_norm": 0.5422080755233765,
|
| 271 |
+
"learning_rate": 6.681918245780636e-06,
|
| 272 |
+
"loss": 0.0898,
|
| 273 |
+
"step": 18000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 7.0,
|
| 277 |
+
"eval_bleu": 0.18638537802601654,
|
| 278 |
+
"eval_brevity_penalty": 0.20073368411830278,
|
| 279 |
+
"eval_length_ratio": 0.3837628106479814,
|
| 280 |
+
"eval_loss": 0.16638144850730896,
|
| 281 |
+
"eval_precision": [
|
| 282 |
+
0.9512606117486455,
|
| 283 |
+
0.9316559946532135,
|
| 284 |
+
0.9199285153866271,
|
| 285 |
+
0.9117109177040873
|
| 286 |
+
],
|
| 287 |
+
"eval_reference_length": 122164,
|
| 288 |
+
"eval_rouge1": 0.5973561321415772,
|
| 289 |
+
"eval_rouge2": 0.5419239447075989,
|
| 290 |
+
"eval_rougeL": 0.5954158036777871,
|
| 291 |
+
"eval_rougeLsum": 0.5954074774518987,
|
| 292 |
+
"eval_runtime": 18.8991,
|
| 293 |
+
"eval_samples_per_second": 26.403,
|
| 294 |
+
"eval_steps_per_second": 1.693,
|
| 295 |
+
"eval_translation_length": 46882,
|
| 296 |
+
"step": 18018
|
| 297 |
+
}
|
| 298 |
+
],
|
| 299 |
+
"logging_steps": 1000,
|
| 300 |
+
"max_steps": 25740,
|
| 301 |
+
"num_input_tokens_seen": 0,
|
| 302 |
+
"num_train_epochs": 10,
|
| 303 |
+
"save_steps": 500,
|
| 304 |
+
"stateful_callbacks": {
|
| 305 |
+
"TrainerControl": {
|
| 306 |
+
"args": {
|
| 307 |
+
"should_epoch_stop": false,
|
| 308 |
+
"should_evaluate": false,
|
| 309 |
+
"should_log": false,
|
| 310 |
+
"should_save": true,
|
| 311 |
+
"should_training_stop": false
|
| 312 |
+
},
|
| 313 |
+
"attributes": {}
|
| 314 |
+
}
|
| 315 |
+
},
|
| 316 |
+
"total_flos": 2.1856563179136e+16,
|
| 317 |
+
"train_batch_size": 4,
|
| 318 |
+
"trial_name": null,
|
| 319 |
+
"trial_params": null
|
| 320 |
+
}
|
checkpoint-18018/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68e795c201e155bf08e2d5a5acbfc1d1c5bb0e489a8c1d705caa5fe3788026b7
|
| 3 |
+
size 5560
|
config.json
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/home/mootez/projects/def-tusharma/mootez/local_models/codet5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"classifier_dropout": 0.0,
|
| 8 |
+
"d_ff": 3072,
|
| 9 |
+
"d_kv": 64,
|
| 10 |
+
"d_model": 768,
|
| 11 |
+
"decoder_start_token_id": 0,
|
| 12 |
+
"dense_act_fn": "relu",
|
| 13 |
+
"dropout_rate": 0.1,
|
| 14 |
+
"eos_token_id": 2,
|
| 15 |
+
"feed_forward_proj": "relu",
|
| 16 |
+
"gradient_checkpointing": false,
|
| 17 |
+
"id2label": {
|
| 18 |
+
"0": "LABEL_0"
|
| 19 |
+
},
|
| 20 |
+
"initializer_factor": 1.0,
|
| 21 |
+
"is_encoder_decoder": true,
|
| 22 |
+
"is_gated_act": false,
|
| 23 |
+
"label2id": {
|
| 24 |
+
"LABEL_0": 0
|
| 25 |
+
},
|
| 26 |
+
"layer_norm_epsilon": 1e-06,
|
| 27 |
+
"merge_layer_index": 5,
|
| 28 |
+
"merger_type": "mean",
|
| 29 |
+
"model_type": "t5",
|
| 30 |
+
"n_positions": 512,
|
| 31 |
+
"num_decoder_layers": 12,
|
| 32 |
+
"num_heads": 12,
|
| 33 |
+
"num_layers": 12,
|
| 34 |
+
"output_past": true,
|
| 35 |
+
"pad_token_id": 0,
|
| 36 |
+
"relative_attention_max_distance": 128,
|
| 37 |
+
"relative_attention_num_buckets": 32,
|
| 38 |
+
"task_specific_params": {
|
| 39 |
+
"summarization": {
|
| 40 |
+
"early_stopping": true,
|
| 41 |
+
"length_penalty": 2.0,
|
| 42 |
+
"max_length": 200,
|
| 43 |
+
"min_length": 30,
|
| 44 |
+
"no_repeat_ngram_size": 3,
|
| 45 |
+
"num_beams": 4,
|
| 46 |
+
"prefix": "summarize: "
|
| 47 |
+
},
|
| 48 |
+
"translation_en_to_de": {
|
| 49 |
+
"early_stopping": true,
|
| 50 |
+
"max_length": 300,
|
| 51 |
+
"num_beams": 4,
|
| 52 |
+
"prefix": "translate English to German: "
|
| 53 |
+
},
|
| 54 |
+
"translation_en_to_fr": {
|
| 55 |
+
"early_stopping": true,
|
| 56 |
+
"max_length": 300,
|
| 57 |
+
"num_beams": 4,
|
| 58 |
+
"prefix": "translate English to French: "
|
| 59 |
+
},
|
| 60 |
+
"translation_en_to_ro": {
|
| 61 |
+
"early_stopping": true,
|
| 62 |
+
"max_length": 300,
|
| 63 |
+
"num_beams": 4,
|
| 64 |
+
"prefix": "translate English to Romanian: "
|
| 65 |
+
}
|
| 66 |
+
},
|
| 67 |
+
"torch_dtype": "float32",
|
| 68 |
+
"transformers_version": "4.48.2",
|
| 69 |
+
"use_cache": true,
|
| 70 |
+
"vocab_size": 32100
|
| 71 |
+
}
|
events.out.tfevents.1751355165.gra5
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:952ebbfd2d4cb09e7df330dbd8fc140a4ead9f9ef6a50347dad9d72d0f97b7bf
|
| 3 |
+
size 19672
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 1,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.2"
|
| 7 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bcf62040e13cb3e83c0bc72a75615abe1dbcf8cc9bcaca08ec263e0552ee355a
|
| 3 |
+
size 891558696
|
results.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"rouge1": 0.6654123223107544,
|
| 3 |
+
"rouge2": 0.6171608247957057,
|
| 4 |
+
"rougeL": 0.6630046753709433,
|
| 5 |
+
"rougeLsum": 0.6629848822601063,
|
| 6 |
+
"bleu": 0.5434842064700386,
|
| 7 |
+
"precision": [
|
| 8 |
+
0.5613548140046501,
|
| 9 |
+
0.5457116117555256,
|
| 10 |
+
0.5368642671543166,
|
| 11 |
+
0.5304965540545652
|
| 12 |
+
],
|
| 13 |
+
"brevity_penalty": 1.0,
|
| 14 |
+
"length_ratio": 1.7547728320779992,
|
| 15 |
+
"translation_length": 399553,
|
| 16 |
+
"reference_length": 227695
|
| 17 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68e795c201e155bf08e2d5a5acbfc1d1c5bb0e489a8c1d705caa5fe3788026b7
|
| 3 |
+
size 5560
|