After 10 epochs
Browse files- config.json +32 -0
- generation_config.json +7 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +692 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "duraad/nep-spell-hft",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MT5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"classifier_dropout": 0.0,
|
| 7 |
+
"d_ff": 1024,
|
| 8 |
+
"d_kv": 64,
|
| 9 |
+
"d_model": 512,
|
| 10 |
+
"decoder_start_token_id": 0,
|
| 11 |
+
"dense_act_fn": "gelu_new",
|
| 12 |
+
"dropout_rate": 0.1,
|
| 13 |
+
"eos_token_id": 1,
|
| 14 |
+
"feed_forward_proj": "gated-gelu",
|
| 15 |
+
"initializer_factor": 1.0,
|
| 16 |
+
"is_encoder_decoder": true,
|
| 17 |
+
"is_gated_act": true,
|
| 18 |
+
"layer_norm_epsilon": 1e-06,
|
| 19 |
+
"model_type": "mt5",
|
| 20 |
+
"num_decoder_layers": 8,
|
| 21 |
+
"num_heads": 6,
|
| 22 |
+
"num_layers": 8,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"relative_attention_max_distance": 128,
|
| 25 |
+
"relative_attention_num_buckets": 32,
|
| 26 |
+
"tie_word_embeddings": false,
|
| 27 |
+
"tokenizer_class": "T5Tokenizer",
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"transformers_version": "4.37.2",
|
| 30 |
+
"use_cache": true,
|
| 31 |
+
"vocab_size": 250112
|
| 32 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.37.2"
|
| 7 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71a88dd7c7f1cbb37952e9b1f0b8f18b3302c3287135d70933705fb66682bbb8
|
| 3 |
+
size 1200729512
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9914f35ec4166caaefa365fc45342de2ff25e5e31820e04a2e7d1cc4b5461c6
|
| 3 |
+
size 2401573893
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:575164422ffb07dce40f4f226304ae580f088580435499d59fc33efa74eb8a4e
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3771f987b89d5be40d08367559c2ba7770e79a7c012a7171d9578f9ad0ee0dbd
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,692 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"eval_steps": 20000,
|
| 6 |
+
"global_step": 101152,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08,
|
| 13 |
+
"learning_rate": 4.980149277433699e-05,
|
| 14 |
+
"loss": 0.0094,
|
| 15 |
+
"step": 1000
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.16,
|
| 19 |
+
"learning_rate": 4.9404478323010964e-05,
|
| 20 |
+
"loss": 0.0102,
|
| 21 |
+
"step": 2000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.24,
|
| 25 |
+
"learning_rate": 4.900746387168493e-05,
|
| 26 |
+
"loss": 0.0098,
|
| 27 |
+
"step": 3000
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.32,
|
| 31 |
+
"learning_rate": 4.86104494203589e-05,
|
| 32 |
+
"loss": 0.0101,
|
| 33 |
+
"step": 4000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.4,
|
| 37 |
+
"learning_rate": 4.821343496903287e-05,
|
| 38 |
+
"loss": 0.0098,
|
| 39 |
+
"step": 5000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.47,
|
| 43 |
+
"learning_rate": 4.781642051770684e-05,
|
| 44 |
+
"loss": 0.0097,
|
| 45 |
+
"step": 6000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.55,
|
| 49 |
+
"learning_rate": 4.741940606638082e-05,
|
| 50 |
+
"loss": 0.0099,
|
| 51 |
+
"step": 7000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.63,
|
| 55 |
+
"learning_rate": 4.702239161505479e-05,
|
| 56 |
+
"loss": 0.01,
|
| 57 |
+
"step": 8000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.71,
|
| 61 |
+
"learning_rate": 4.6625377163728765e-05,
|
| 62 |
+
"loss": 0.01,
|
| 63 |
+
"step": 9000
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.79,
|
| 67 |
+
"learning_rate": 4.6228362712402734e-05,
|
| 68 |
+
"loss": 0.0119,
|
| 69 |
+
"step": 10000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.87,
|
| 73 |
+
"learning_rate": 4.5831348261076704e-05,
|
| 74 |
+
"loss": 0.0095,
|
| 75 |
+
"step": 11000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.95,
|
| 79 |
+
"learning_rate": 4.543433380975068e-05,
|
| 80 |
+
"loss": 0.0097,
|
| 81 |
+
"step": 12000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.03,
|
| 85 |
+
"learning_rate": 4.503731935842465e-05,
|
| 86 |
+
"loss": 0.0095,
|
| 87 |
+
"step": 13000
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.11,
|
| 91 |
+
"learning_rate": 4.464030490709862e-05,
|
| 92 |
+
"loss": 0.0089,
|
| 93 |
+
"step": 14000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.19,
|
| 97 |
+
"learning_rate": 4.424329045577259e-05,
|
| 98 |
+
"loss": 0.0088,
|
| 99 |
+
"step": 15000
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.27,
|
| 103 |
+
"learning_rate": 4.3846276004446566e-05,
|
| 104 |
+
"loss": 0.0088,
|
| 105 |
+
"step": 16000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.34,
|
| 109 |
+
"learning_rate": 4.3449261553120535e-05,
|
| 110 |
+
"loss": 0.0085,
|
| 111 |
+
"step": 17000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 1.42,
|
| 115 |
+
"learning_rate": 4.3052247101794505e-05,
|
| 116 |
+
"loss": 0.0087,
|
| 117 |
+
"step": 18000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 1.5,
|
| 121 |
+
"learning_rate": 4.265523265046848e-05,
|
| 122 |
+
"loss": 0.009,
|
| 123 |
+
"step": 19000
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.58,
|
| 127 |
+
"learning_rate": 4.225821819914245e-05,
|
| 128 |
+
"loss": 0.0087,
|
| 129 |
+
"step": 20000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.58,
|
| 133 |
+
"eval_accuracy": 0.5290210343191523,
|
| 134 |
+
"eval_exact_match": 0.5290210343191523,
|
| 135 |
+
"eval_f1": 0.531319521324266,
|
| 136 |
+
"eval_loss": 0.005351942032575607,
|
| 137 |
+
"eval_precision": 0.5358216036691444,
|
| 138 |
+
"eval_recall": 0.5290210343191523,
|
| 139 |
+
"eval_runtime": 356.516,
|
| 140 |
+
"eval_samples_per_second": 17.736,
|
| 141 |
+
"eval_steps_per_second": 8.869,
|
| 142 |
+
"step": 20000
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 1.66,
|
| 146 |
+
"learning_rate": 4.186120374781643e-05,
|
| 147 |
+
"loss": 0.0082,
|
| 148 |
+
"step": 21000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 1.74,
|
| 152 |
+
"learning_rate": 4.14641892964904e-05,
|
| 153 |
+
"loss": 0.0084,
|
| 154 |
+
"step": 22000
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.82,
|
| 158 |
+
"learning_rate": 4.1067174845164366e-05,
|
| 159 |
+
"loss": 0.0082,
|
| 160 |
+
"step": 23000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.9,
|
| 164 |
+
"learning_rate": 4.0670160393838336e-05,
|
| 165 |
+
"loss": 0.0084,
|
| 166 |
+
"step": 24000
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 1.98,
|
| 170 |
+
"learning_rate": 4.0273145942512306e-05,
|
| 171 |
+
"loss": 0.0084,
|
| 172 |
+
"step": 25000
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.06,
|
| 176 |
+
"learning_rate": 3.987613149118628e-05,
|
| 177 |
+
"loss": 0.008,
|
| 178 |
+
"step": 26000
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 2.14,
|
| 182 |
+
"learning_rate": 3.947911703986025e-05,
|
| 183 |
+
"loss": 0.0075,
|
| 184 |
+
"step": 27000
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 2.21,
|
| 188 |
+
"learning_rate": 3.908210258853423e-05,
|
| 189 |
+
"loss": 0.0078,
|
| 190 |
+
"step": 28000
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 2.29,
|
| 194 |
+
"learning_rate": 3.86850881372082e-05,
|
| 195 |
+
"loss": 0.008,
|
| 196 |
+
"step": 29000
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 2.37,
|
| 200 |
+
"learning_rate": 3.828807368588217e-05,
|
| 201 |
+
"loss": 0.0075,
|
| 202 |
+
"step": 30000
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 2.45,
|
| 206 |
+
"learning_rate": 3.7891059234556144e-05,
|
| 207 |
+
"loss": 0.0077,
|
| 208 |
+
"step": 31000
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 2.53,
|
| 212 |
+
"learning_rate": 3.749404478323011e-05,
|
| 213 |
+
"loss": 0.0074,
|
| 214 |
+
"step": 32000
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 2.61,
|
| 218 |
+
"learning_rate": 3.709703033190408e-05,
|
| 219 |
+
"loss": 0.0078,
|
| 220 |
+
"step": 33000
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 2.69,
|
| 224 |
+
"learning_rate": 3.670001588057805e-05,
|
| 225 |
+
"loss": 0.008,
|
| 226 |
+
"step": 34000
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 2.77,
|
| 230 |
+
"learning_rate": 3.630300142925203e-05,
|
| 231 |
+
"loss": 0.0073,
|
| 232 |
+
"step": 35000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 2.85,
|
| 236 |
+
"learning_rate": 3.5905986977926e-05,
|
| 237 |
+
"loss": 0.0076,
|
| 238 |
+
"step": 36000
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 2.93,
|
| 242 |
+
"learning_rate": 3.550897252659997e-05,
|
| 243 |
+
"loss": 0.0076,
|
| 244 |
+
"step": 37000
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 3.01,
|
| 248 |
+
"learning_rate": 3.5111958075273945e-05,
|
| 249 |
+
"loss": 0.0078,
|
| 250 |
+
"step": 38000
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 3.08,
|
| 254 |
+
"learning_rate": 3.4714943623947914e-05,
|
| 255 |
+
"loss": 0.007,
|
| 256 |
+
"step": 39000
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 3.16,
|
| 260 |
+
"learning_rate": 3.431792917262189e-05,
|
| 261 |
+
"loss": 0.0065,
|
| 262 |
+
"step": 40000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 3.16,
|
| 266 |
+
"eval_accuracy": 0.5473667562865728,
|
| 267 |
+
"eval_exact_match": 0.5473667562865728,
|
| 268 |
+
"eval_f1": 0.5494543729242448,
|
| 269 |
+
"eval_loss": 0.005287344101816416,
|
| 270 |
+
"eval_precision": 0.5535347145342401,
|
| 271 |
+
"eval_recall": 0.5473667562865728,
|
| 272 |
+
"eval_runtime": 354.3494,
|
| 273 |
+
"eval_samples_per_second": 17.844,
|
| 274 |
+
"eval_steps_per_second": 8.923,
|
| 275 |
+
"step": 40000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 3.24,
|
| 279 |
+
"learning_rate": 3.3920914721295854e-05,
|
| 280 |
+
"loss": 0.007,
|
| 281 |
+
"step": 41000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.32,
|
| 285 |
+
"learning_rate": 3.352390026996983e-05,
|
| 286 |
+
"loss": 0.0071,
|
| 287 |
+
"step": 42000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 3.4,
|
| 291 |
+
"learning_rate": 3.31268858186438e-05,
|
| 292 |
+
"loss": 0.0074,
|
| 293 |
+
"step": 43000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 3.48,
|
| 297 |
+
"learning_rate": 3.272987136731777e-05,
|
| 298 |
+
"loss": 0.0068,
|
| 299 |
+
"step": 44000
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 3.56,
|
| 303 |
+
"learning_rate": 3.2332856915991746e-05,
|
| 304 |
+
"loss": 0.0067,
|
| 305 |
+
"step": 45000
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.64,
|
| 309 |
+
"learning_rate": 3.1935842464665715e-05,
|
| 310 |
+
"loss": 0.0071,
|
| 311 |
+
"step": 46000
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.72,
|
| 315 |
+
"learning_rate": 3.153882801333969e-05,
|
| 316 |
+
"loss": 0.0067,
|
| 317 |
+
"step": 47000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 3.8,
|
| 321 |
+
"learning_rate": 3.114181356201366e-05,
|
| 322 |
+
"loss": 0.0071,
|
| 323 |
+
"step": 48000
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 3.88,
|
| 327 |
+
"learning_rate": 3.074479911068763e-05,
|
| 328 |
+
"loss": 0.0073,
|
| 329 |
+
"step": 49000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.95,
|
| 333 |
+
"learning_rate": 3.03477846593616e-05,
|
| 334 |
+
"loss": 0.0067,
|
| 335 |
+
"step": 50000
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 4.03,
|
| 339 |
+
"learning_rate": 2.9950770208035574e-05,
|
| 340 |
+
"loss": 0.0067,
|
| 341 |
+
"step": 51000
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 4.11,
|
| 345 |
+
"learning_rate": 2.9553755756709543e-05,
|
| 346 |
+
"loss": 0.0064,
|
| 347 |
+
"step": 52000
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 4.19,
|
| 351 |
+
"learning_rate": 2.9156741305383516e-05,
|
| 352 |
+
"loss": 0.0062,
|
| 353 |
+
"step": 53000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 4.27,
|
| 357 |
+
"learning_rate": 2.875972685405749e-05,
|
| 358 |
+
"loss": 0.0066,
|
| 359 |
+
"step": 54000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 4.35,
|
| 363 |
+
"learning_rate": 2.8362712402731462e-05,
|
| 364 |
+
"loss": 0.0062,
|
| 365 |
+
"step": 55000
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 4.43,
|
| 369 |
+
"learning_rate": 2.7965697951405435e-05,
|
| 370 |
+
"loss": 0.0065,
|
| 371 |
+
"step": 56000
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 4.51,
|
| 375 |
+
"learning_rate": 2.7568683500079405e-05,
|
| 376 |
+
"loss": 0.0066,
|
| 377 |
+
"step": 57000
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 4.59,
|
| 381 |
+
"learning_rate": 2.7171669048753375e-05,
|
| 382 |
+
"loss": 0.0063,
|
| 383 |
+
"step": 58000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 4.67,
|
| 387 |
+
"learning_rate": 2.6774654597427344e-05,
|
| 388 |
+
"loss": 0.0064,
|
| 389 |
+
"step": 59000
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 4.75,
|
| 393 |
+
"learning_rate": 2.6377640146101317e-05,
|
| 394 |
+
"loss": 0.006,
|
| 395 |
+
"step": 60000
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 4.75,
|
| 399 |
+
"eval_accuracy": 0.5701407559702673,
|
| 400 |
+
"eval_exact_match": 0.5701407559702673,
|
| 401 |
+
"eval_f1": 0.5724919605672413,
|
| 402 |
+
"eval_loss": 0.0050529008731245995,
|
| 403 |
+
"eval_precision": 0.5770994780958406,
|
| 404 |
+
"eval_recall": 0.5701407559702673,
|
| 405 |
+
"eval_runtime": 354.5246,
|
| 406 |
+
"eval_samples_per_second": 17.835,
|
| 407 |
+
"eval_steps_per_second": 8.919,
|
| 408 |
+
"step": 60000
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 4.82,
|
| 412 |
+
"learning_rate": 2.598062569477529e-05,
|
| 413 |
+
"loss": 0.0065,
|
| 414 |
+
"step": 61000
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 4.9,
|
| 418 |
+
"learning_rate": 2.5583611243449263e-05,
|
| 419 |
+
"loss": 0.0063,
|
| 420 |
+
"step": 62000
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 4.98,
|
| 424 |
+
"learning_rate": 2.5186596792123236e-05,
|
| 425 |
+
"loss": 0.0068,
|
| 426 |
+
"step": 63000
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 5.06,
|
| 430 |
+
"learning_rate": 2.4789582340797206e-05,
|
| 431 |
+
"loss": 0.006,
|
| 432 |
+
"step": 64000
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 5.14,
|
| 436 |
+
"learning_rate": 2.4392567889471176e-05,
|
| 437 |
+
"loss": 0.006,
|
| 438 |
+
"step": 65000
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 5.22,
|
| 442 |
+
"learning_rate": 2.399555343814515e-05,
|
| 443 |
+
"loss": 0.0058,
|
| 444 |
+
"step": 66000
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 5.3,
|
| 448 |
+
"learning_rate": 2.359853898681912e-05,
|
| 449 |
+
"loss": 0.0056,
|
| 450 |
+
"step": 67000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 5.38,
|
| 454 |
+
"learning_rate": 2.3201524535493095e-05,
|
| 455 |
+
"loss": 0.0057,
|
| 456 |
+
"step": 68000
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 5.46,
|
| 460 |
+
"learning_rate": 2.2804510084167068e-05,
|
| 461 |
+
"loss": 0.0059,
|
| 462 |
+
"step": 69000
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 5.54,
|
| 466 |
+
"learning_rate": 2.2407495632841037e-05,
|
| 467 |
+
"loss": 0.0057,
|
| 468 |
+
"step": 70000
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 5.62,
|
| 472 |
+
"learning_rate": 2.2010481181515007e-05,
|
| 473 |
+
"loss": 0.0058,
|
| 474 |
+
"step": 71000
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 5.69,
|
| 478 |
+
"learning_rate": 2.161346673018898e-05,
|
| 479 |
+
"loss": 0.0059,
|
| 480 |
+
"step": 72000
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 5.77,
|
| 484 |
+
"learning_rate": 2.1216452278862953e-05,
|
| 485 |
+
"loss": 0.0059,
|
| 486 |
+
"step": 73000
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 5.85,
|
| 490 |
+
"learning_rate": 2.0819437827536923e-05,
|
| 491 |
+
"loss": 0.006,
|
| 492 |
+
"step": 74000
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 5.93,
|
| 496 |
+
"learning_rate": 2.0422423376210896e-05,
|
| 497 |
+
"loss": 0.0056,
|
| 498 |
+
"step": 75000
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 6.01,
|
| 502 |
+
"learning_rate": 2.002540892488487e-05,
|
| 503 |
+
"loss": 0.0058,
|
| 504 |
+
"step": 76000
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 6.09,
|
| 508 |
+
"learning_rate": 1.9628394473558838e-05,
|
| 509 |
+
"loss": 0.0054,
|
| 510 |
+
"step": 77000
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 6.17,
|
| 514 |
+
"learning_rate": 1.9231380022232808e-05,
|
| 515 |
+
"loss": 0.0061,
|
| 516 |
+
"step": 78000
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 6.25,
|
| 520 |
+
"learning_rate": 1.883436557090678e-05,
|
| 521 |
+
"loss": 0.0057,
|
| 522 |
+
"step": 79000
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 6.33,
|
| 526 |
+
"learning_rate": 1.8437351119580754e-05,
|
| 527 |
+
"loss": 0.0053,
|
| 528 |
+
"step": 80000
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 6.33,
|
| 532 |
+
"eval_accuracy": 0.5827929780167642,
|
| 533 |
+
"eval_exact_match": 0.5827929780167642,
|
| 534 |
+
"eval_f1": 0.5851441826137382,
|
| 535 |
+
"eval_loss": 0.004922935273498297,
|
| 536 |
+
"eval_precision": 0.5897517001423375,
|
| 537 |
+
"eval_recall": 0.5827929780167642,
|
| 538 |
+
"eval_runtime": 354.2336,
|
| 539 |
+
"eval_samples_per_second": 17.85,
|
| 540 |
+
"eval_steps_per_second": 8.926,
|
| 541 |
+
"step": 80000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 6.41,
|
| 545 |
+
"learning_rate": 1.8040336668254727e-05,
|
| 546 |
+
"loss": 0.0054,
|
| 547 |
+
"step": 81000
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 6.49,
|
| 551 |
+
"learning_rate": 1.76433222169287e-05,
|
| 552 |
+
"loss": 0.0054,
|
| 553 |
+
"step": 82000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 6.56,
|
| 557 |
+
"learning_rate": 1.7246307765602666e-05,
|
| 558 |
+
"loss": 0.0056,
|
| 559 |
+
"step": 83000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 6.64,
|
| 563 |
+
"learning_rate": 1.684929331427664e-05,
|
| 564 |
+
"loss": 0.0055,
|
| 565 |
+
"step": 84000
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 6.72,
|
| 569 |
+
"learning_rate": 1.6452278862950612e-05,
|
| 570 |
+
"loss": 0.0051,
|
| 571 |
+
"step": 85000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 6.8,
|
| 575 |
+
"learning_rate": 1.6055264411624585e-05,
|
| 576 |
+
"loss": 0.0057,
|
| 577 |
+
"step": 86000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 6.88,
|
| 581 |
+
"learning_rate": 1.5658249960298555e-05,
|
| 582 |
+
"loss": 0.0053,
|
| 583 |
+
"step": 87000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 6.96,
|
| 587 |
+
"learning_rate": 1.5261235508972528e-05,
|
| 588 |
+
"loss": 0.0053,
|
| 589 |
+
"step": 88000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 7.04,
|
| 593 |
+
"learning_rate": 1.4864221057646499e-05,
|
| 594 |
+
"loss": 0.0049,
|
| 595 |
+
"step": 89000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 7.12,
|
| 599 |
+
"learning_rate": 1.4467206606320472e-05,
|
| 600 |
+
"loss": 0.0054,
|
| 601 |
+
"step": 90000
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 7.2,
|
| 605 |
+
"learning_rate": 1.4070192154994443e-05,
|
| 606 |
+
"loss": 0.0052,
|
| 607 |
+
"step": 91000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 7.28,
|
| 611 |
+
"learning_rate": 1.3673177703668413e-05,
|
| 612 |
+
"loss": 0.0052,
|
| 613 |
+
"step": 92000
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 7.36,
|
| 617 |
+
"learning_rate": 1.3276163252342386e-05,
|
| 618 |
+
"loss": 0.0053,
|
| 619 |
+
"step": 93000
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 7.43,
|
| 623 |
+
"learning_rate": 1.2879148801016357e-05,
|
| 624 |
+
"loss": 0.005,
|
| 625 |
+
"step": 94000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 7.51,
|
| 629 |
+
"learning_rate": 1.2482134349690329e-05,
|
| 630 |
+
"loss": 0.0054,
|
| 631 |
+
"step": 95000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 7.59,
|
| 635 |
+
"learning_rate": 1.2085119898364302e-05,
|
| 636 |
+
"loss": 0.0052,
|
| 637 |
+
"step": 96000
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 7.67,
|
| 641 |
+
"learning_rate": 1.1688105447038273e-05,
|
| 642 |
+
"loss": 0.0051,
|
| 643 |
+
"step": 97000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 7.75,
|
| 647 |
+
"learning_rate": 1.1291090995712244e-05,
|
| 648 |
+
"loss": 0.005,
|
| 649 |
+
"step": 98000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 7.83,
|
| 653 |
+
"learning_rate": 1.0894076544386216e-05,
|
| 654 |
+
"loss": 0.005,
|
| 655 |
+
"step": 99000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 7.91,
|
| 659 |
+
"learning_rate": 1.0497062093060187e-05,
|
| 660 |
+
"loss": 0.005,
|
| 661 |
+
"step": 100000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 7.91,
|
| 665 |
+
"eval_accuracy": 0.5921239917760557,
|
| 666 |
+
"eval_exact_match": 0.5921239917760557,
|
| 667 |
+
"eval_f1": 0.594369761189309,
|
| 668 |
+
"eval_loss": 0.004835059866309166,
|
| 669 |
+
"eval_precision": 0.5987664083504666,
|
| 670 |
+
"eval_recall": 0.5921239917760557,
|
| 671 |
+
"eval_runtime": 354.4854,
|
| 672 |
+
"eval_samples_per_second": 17.837,
|
| 673 |
+
"eval_steps_per_second": 8.92,
|
| 674 |
+
"step": 100000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 7.99,
|
| 678 |
+
"learning_rate": 1.010004764173416e-05,
|
| 679 |
+
"loss": 0.005,
|
| 680 |
+
"step": 101000
|
| 681 |
+
}
|
| 682 |
+
],
|
| 683 |
+
"logging_steps": 1000,
|
| 684 |
+
"max_steps": 126440,
|
| 685 |
+
"num_input_tokens_seen": 0,
|
| 686 |
+
"num_train_epochs": 10,
|
| 687 |
+
"save_steps": 500,
|
| 688 |
+
"total_flos": 1.0696817746378752e+17,
|
| 689 |
+
"train_batch_size": 2,
|
| 690 |
+
"trial_name": null,
|
| 691 |
+
"trial_params": null
|
| 692 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c55e3399602b93c4ba608798a65f62e367a8d7bc1a83bdefd9ed7b7a6a55170
|
| 3 |
+
size 4219
|