Upload folder using huggingface_hub
Browse files- config.json +75 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +830 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/bart-large",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"add_bias_logits": false,
|
| 6 |
+
"add_final_layer_norm": false,
|
| 7 |
+
"architectures": [
|
| 8 |
+
"BartForSequenceClassification"
|
| 9 |
+
],
|
| 10 |
+
"attention_dropout": 0.1,
|
| 11 |
+
"bos_token_id": 0,
|
| 12 |
+
"classif_dropout": 0.1,
|
| 13 |
+
"classifier_dropout": 0.0,
|
| 14 |
+
"d_model": 1024,
|
| 15 |
+
"decoder_attention_heads": 16,
|
| 16 |
+
"decoder_ffn_dim": 4096,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 12,
|
| 19 |
+
"decoder_start_token_id": 2,
|
| 20 |
+
"dropout": 0.1,
|
| 21 |
+
"early_stopping": true,
|
| 22 |
+
"encoder_attention_heads": 16,
|
| 23 |
+
"encoder_ffn_dim": 4096,
|
| 24 |
+
"encoder_layerdrop": 0.0,
|
| 25 |
+
"encoder_layers": 12,
|
| 26 |
+
"eos_token_id": 2,
|
| 27 |
+
"forced_bos_token_id": 0,
|
| 28 |
+
"forced_eos_token_id": 2,
|
| 29 |
+
"gradient_checkpointing": false,
|
| 30 |
+
"id2label": {
|
| 31 |
+
"0": "LABEL_0",
|
| 32 |
+
"1": "LABEL_1",
|
| 33 |
+
"2": "LABEL_2"
|
| 34 |
+
},
|
| 35 |
+
"init_std": 0.02,
|
| 36 |
+
"is_encoder_decoder": true,
|
| 37 |
+
"label2id": {
|
| 38 |
+
"LABEL_0": 0,
|
| 39 |
+
"LABEL_1": 1,
|
| 40 |
+
"LABEL_2": 2
|
| 41 |
+
},
|
| 42 |
+
"max_position_embeddings": 1024,
|
| 43 |
+
"model_type": "bart",
|
| 44 |
+
"no_repeat_ngram_size": 3,
|
| 45 |
+
"normalize_before": false,
|
| 46 |
+
"num_beams": 4,
|
| 47 |
+
"num_hidden_layers": 12,
|
| 48 |
+
"pad_token_id": 1,
|
| 49 |
+
"problem_type": "single_label_classification",
|
| 50 |
+
"scale_embedding": false,
|
| 51 |
+
"task_specific_params": {
|
| 52 |
+
"summarization": {
|
| 53 |
+
"length_penalty": 1.0,
|
| 54 |
+
"max_length": 128,
|
| 55 |
+
"min_length": 12,
|
| 56 |
+
"num_beams": 4
|
| 57 |
+
},
|
| 58 |
+
"summarization_cnn": {
|
| 59 |
+
"length_penalty": 2.0,
|
| 60 |
+
"max_length": 142,
|
| 61 |
+
"min_length": 56,
|
| 62 |
+
"num_beams": 4
|
| 63 |
+
},
|
| 64 |
+
"summarization_xsum": {
|
| 65 |
+
"length_penalty": 1.0,
|
| 66 |
+
"max_length": 62,
|
| 67 |
+
"min_length": 11,
|
| 68 |
+
"num_beams": 6
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
"torch_dtype": "float32",
|
| 72 |
+
"transformers_version": "4.38.2",
|
| 73 |
+
"use_cache": true,
|
| 74 |
+
"vocab_size": 50265
|
| 75 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ea8a9bfa96dc6ba6e9857fc52e923b9177bfbde5ac6adf0e711ed56438502e5
|
| 3 |
+
size 1629436964
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85d4e92201a3f6a0ccc7ac152dcf5666f224a02bc881fa3af5b316246dabd49c
|
| 3 |
+
size 3259184731
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:43cf595b9820e159a0a94c69432baaecbdf7c55cd5a206271129e17ef7b0697c
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be0acaa0c2956be0b218303bd552582151ffb6ce7e7c85cf55e7de8ae4ec7b1c
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,830 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 18.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 9522,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.19,
|
| 13 |
+
"grad_norm": 6.645576477050781,
|
| 14 |
+
"learning_rate": 1.9810964083175805e-05,
|
| 15 |
+
"loss": 0.5588,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.38,
|
| 20 |
+
"grad_norm": 4.662398815155029,
|
| 21 |
+
"learning_rate": 1.962192816635161e-05,
|
| 22 |
+
"loss": 0.3737,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.57,
|
| 27 |
+
"grad_norm": 5.282985687255859,
|
| 28 |
+
"learning_rate": 1.9432892249527412e-05,
|
| 29 |
+
"loss": 0.2863,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.76,
|
| 34 |
+
"grad_norm": 1.5854848623275757,
|
| 35 |
+
"learning_rate": 1.9243856332703215e-05,
|
| 36 |
+
"loss": 0.323,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.95,
|
| 41 |
+
"grad_norm": 19.09914779663086,
|
| 42 |
+
"learning_rate": 1.905482041587902e-05,
|
| 43 |
+
"loss": 0.2956,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1.0,
|
| 48 |
+
"eval_loss": 0.20477035641670227,
|
| 49 |
+
"eval_runtime": 4.245,
|
| 50 |
+
"eval_samples_per_second": 91.638,
|
| 51 |
+
"eval_steps_per_second": 6.596,
|
| 52 |
+
"step": 529
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.13,
|
| 56 |
+
"grad_norm": 26.59105682373047,
|
| 57 |
+
"learning_rate": 1.8865784499054822e-05,
|
| 58 |
+
"loss": 0.2269,
|
| 59 |
+
"step": 600
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.32,
|
| 63 |
+
"grad_norm": 9.394390106201172,
|
| 64 |
+
"learning_rate": 1.8676748582230626e-05,
|
| 65 |
+
"loss": 0.2393,
|
| 66 |
+
"step": 700
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.51,
|
| 70 |
+
"grad_norm": 13.779593467712402,
|
| 71 |
+
"learning_rate": 1.848771266540643e-05,
|
| 72 |
+
"loss": 0.2135,
|
| 73 |
+
"step": 800
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 1.7,
|
| 77 |
+
"grad_norm": 0.6053704619407654,
|
| 78 |
+
"learning_rate": 1.8298676748582232e-05,
|
| 79 |
+
"loss": 0.2241,
|
| 80 |
+
"step": 900
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.89,
|
| 84 |
+
"grad_norm": 16.190296173095703,
|
| 85 |
+
"learning_rate": 1.8109640831758036e-05,
|
| 86 |
+
"loss": 0.2355,
|
| 87 |
+
"step": 1000
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 2.0,
|
| 91 |
+
"eval_loss": 0.21921110153198242,
|
| 92 |
+
"eval_runtime": 4.2197,
|
| 93 |
+
"eval_samples_per_second": 92.187,
|
| 94 |
+
"eval_steps_per_second": 6.636,
|
| 95 |
+
"step": 1058
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.08,
|
| 99 |
+
"grad_norm": 0.10688259452581406,
|
| 100 |
+
"learning_rate": 1.792060491493384e-05,
|
| 101 |
+
"loss": 0.1909,
|
| 102 |
+
"step": 1100
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.27,
|
| 106 |
+
"grad_norm": 29.609710693359375,
|
| 107 |
+
"learning_rate": 1.7731568998109643e-05,
|
| 108 |
+
"loss": 0.1637,
|
| 109 |
+
"step": 1200
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 2.46,
|
| 113 |
+
"grad_norm": 0.16163112223148346,
|
| 114 |
+
"learning_rate": 1.7542533081285446e-05,
|
| 115 |
+
"loss": 0.1828,
|
| 116 |
+
"step": 1300
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.65,
|
| 120 |
+
"grad_norm": 2.127821207046509,
|
| 121 |
+
"learning_rate": 1.735349716446125e-05,
|
| 122 |
+
"loss": 0.1827,
|
| 123 |
+
"step": 1400
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.84,
|
| 127 |
+
"grad_norm": 0.030160142108798027,
|
| 128 |
+
"learning_rate": 1.7164461247637053e-05,
|
| 129 |
+
"loss": 0.1528,
|
| 130 |
+
"step": 1500
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 3.0,
|
| 134 |
+
"eval_loss": 0.2590792179107666,
|
| 135 |
+
"eval_runtime": 4.2371,
|
| 136 |
+
"eval_samples_per_second": 91.808,
|
| 137 |
+
"eval_steps_per_second": 6.608,
|
| 138 |
+
"step": 1587
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 3.02,
|
| 142 |
+
"grad_norm": 0.0637928918004036,
|
| 143 |
+
"learning_rate": 1.6975425330812856e-05,
|
| 144 |
+
"loss": 0.1574,
|
| 145 |
+
"step": 1600
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 3.21,
|
| 149 |
+
"grad_norm": 0.01419603731483221,
|
| 150 |
+
"learning_rate": 1.678638941398866e-05,
|
| 151 |
+
"loss": 0.0918,
|
| 152 |
+
"step": 1700
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.4,
|
| 156 |
+
"grad_norm": 9.31396770477295,
|
| 157 |
+
"learning_rate": 1.6597353497164463e-05,
|
| 158 |
+
"loss": 0.1175,
|
| 159 |
+
"step": 1800
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 3.59,
|
| 163 |
+
"grad_norm": 0.023506687954068184,
|
| 164 |
+
"learning_rate": 1.6408317580340267e-05,
|
| 165 |
+
"loss": 0.1094,
|
| 166 |
+
"step": 1900
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.78,
|
| 170 |
+
"grad_norm": 13.4344482421875,
|
| 171 |
+
"learning_rate": 1.621928166351607e-05,
|
| 172 |
+
"loss": 0.1103,
|
| 173 |
+
"step": 2000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 3.97,
|
| 177 |
+
"grad_norm": 30.01926040649414,
|
| 178 |
+
"learning_rate": 1.6030245746691873e-05,
|
| 179 |
+
"loss": 0.1398,
|
| 180 |
+
"step": 2100
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 4.0,
|
| 184 |
+
"eval_loss": 0.33959057927131653,
|
| 185 |
+
"eval_runtime": 4.2467,
|
| 186 |
+
"eval_samples_per_second": 91.601,
|
| 187 |
+
"eval_steps_per_second": 6.593,
|
| 188 |
+
"step": 2116
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 4.16,
|
| 192 |
+
"grad_norm": 0.12888406217098236,
|
| 193 |
+
"learning_rate": 1.5841209829867677e-05,
|
| 194 |
+
"loss": 0.0536,
|
| 195 |
+
"step": 2200
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 4.35,
|
| 199 |
+
"grad_norm": 0.06800971180200577,
|
| 200 |
+
"learning_rate": 1.565217391304348e-05,
|
| 201 |
+
"loss": 0.1039,
|
| 202 |
+
"step": 2300
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 4.54,
|
| 206 |
+
"grad_norm": 2.495684862136841,
|
| 207 |
+
"learning_rate": 1.5463137996219284e-05,
|
| 208 |
+
"loss": 0.0886,
|
| 209 |
+
"step": 2400
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 4.73,
|
| 213 |
+
"grad_norm": 3.7909159660339355,
|
| 214 |
+
"learning_rate": 1.5274102079395087e-05,
|
| 215 |
+
"loss": 0.0766,
|
| 216 |
+
"step": 2500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 4.91,
|
| 220 |
+
"grad_norm": 0.025901077315211296,
|
| 221 |
+
"learning_rate": 1.5085066162570889e-05,
|
| 222 |
+
"loss": 0.0916,
|
| 223 |
+
"step": 2600
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 5.0,
|
| 227 |
+
"eval_loss": 0.3040069341659546,
|
| 228 |
+
"eval_runtime": 4.2484,
|
| 229 |
+
"eval_samples_per_second": 91.564,
|
| 230 |
+
"eval_steps_per_second": 6.591,
|
| 231 |
+
"step": 2645
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 5.1,
|
| 235 |
+
"grad_norm": 0.011547481641173363,
|
| 236 |
+
"learning_rate": 1.4896030245746694e-05,
|
| 237 |
+
"loss": 0.0644,
|
| 238 |
+
"step": 2700
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 5.29,
|
| 242 |
+
"grad_norm": 0.026049258187413216,
|
| 243 |
+
"learning_rate": 1.4706994328922497e-05,
|
| 244 |
+
"loss": 0.0542,
|
| 245 |
+
"step": 2800
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 5.48,
|
| 249 |
+
"grad_norm": 0.35288381576538086,
|
| 250 |
+
"learning_rate": 1.45179584120983e-05,
|
| 251 |
+
"loss": 0.0512,
|
| 252 |
+
"step": 2900
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 5.67,
|
| 256 |
+
"grad_norm": 0.023430563509464264,
|
| 257 |
+
"learning_rate": 1.4328922495274103e-05,
|
| 258 |
+
"loss": 0.0491,
|
| 259 |
+
"step": 3000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 5.86,
|
| 263 |
+
"grad_norm": 0.0032355256844311953,
|
| 264 |
+
"learning_rate": 1.4139886578449906e-05,
|
| 265 |
+
"loss": 0.0326,
|
| 266 |
+
"step": 3100
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 6.0,
|
| 270 |
+
"eval_loss": 0.3605019450187683,
|
| 271 |
+
"eval_runtime": 4.2462,
|
| 272 |
+
"eval_samples_per_second": 91.611,
|
| 273 |
+
"eval_steps_per_second": 6.594,
|
| 274 |
+
"step": 3174
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 6.05,
|
| 278 |
+
"grad_norm": 0.1259104311466217,
|
| 279 |
+
"learning_rate": 1.395085066162571e-05,
|
| 280 |
+
"loss": 0.0665,
|
| 281 |
+
"step": 3200
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 6.24,
|
| 285 |
+
"grad_norm": 0.08823427557945251,
|
| 286 |
+
"learning_rate": 1.3761814744801514e-05,
|
| 287 |
+
"loss": 0.0252,
|
| 288 |
+
"step": 3300
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 6.43,
|
| 292 |
+
"grad_norm": 101.22339630126953,
|
| 293 |
+
"learning_rate": 1.3572778827977318e-05,
|
| 294 |
+
"loss": 0.0416,
|
| 295 |
+
"step": 3400
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 6.62,
|
| 299 |
+
"grad_norm": 0.033835213631391525,
|
| 300 |
+
"learning_rate": 1.338374291115312e-05,
|
| 301 |
+
"loss": 0.0465,
|
| 302 |
+
"step": 3500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 6.81,
|
| 306 |
+
"grad_norm": 0.15451325476169586,
|
| 307 |
+
"learning_rate": 1.3194706994328923e-05,
|
| 308 |
+
"loss": 0.0358,
|
| 309 |
+
"step": 3600
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 6.99,
|
| 313 |
+
"grad_norm": 0.0031314522493630648,
|
| 314 |
+
"learning_rate": 1.3005671077504726e-05,
|
| 315 |
+
"loss": 0.0344,
|
| 316 |
+
"step": 3700
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 7.0,
|
| 320 |
+
"eval_loss": 0.41286924481391907,
|
| 321 |
+
"eval_runtime": 4.2519,
|
| 322 |
+
"eval_samples_per_second": 91.488,
|
| 323 |
+
"eval_steps_per_second": 6.585,
|
| 324 |
+
"step": 3703
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 7.18,
|
| 328 |
+
"grad_norm": 0.0014451502356678247,
|
| 329 |
+
"learning_rate": 1.281663516068053e-05,
|
| 330 |
+
"loss": 0.0148,
|
| 331 |
+
"step": 3800
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 7.37,
|
| 335 |
+
"grad_norm": 17.679460525512695,
|
| 336 |
+
"learning_rate": 1.2627599243856335e-05,
|
| 337 |
+
"loss": 0.0309,
|
| 338 |
+
"step": 3900
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 7.56,
|
| 342 |
+
"grad_norm": 0.03577155992388725,
|
| 343 |
+
"learning_rate": 1.2438563327032138e-05,
|
| 344 |
+
"loss": 0.0235,
|
| 345 |
+
"step": 4000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 7.75,
|
| 349 |
+
"grad_norm": 28.639484405517578,
|
| 350 |
+
"learning_rate": 1.224952741020794e-05,
|
| 351 |
+
"loss": 0.0296,
|
| 352 |
+
"step": 4100
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 7.94,
|
| 356 |
+
"grad_norm": 0.00975144561380148,
|
| 357 |
+
"learning_rate": 1.2060491493383744e-05,
|
| 358 |
+
"loss": 0.0218,
|
| 359 |
+
"step": 4200
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 8.0,
|
| 363 |
+
"eval_loss": 0.3556581735610962,
|
| 364 |
+
"eval_runtime": 4.245,
|
| 365 |
+
"eval_samples_per_second": 91.637,
|
| 366 |
+
"eval_steps_per_second": 6.596,
|
| 367 |
+
"step": 4232
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 8.13,
|
| 371 |
+
"grad_norm": 0.019052820280194283,
|
| 372 |
+
"learning_rate": 1.1871455576559547e-05,
|
| 373 |
+
"loss": 0.0192,
|
| 374 |
+
"step": 4300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 8.32,
|
| 378 |
+
"grad_norm": 12.919537544250488,
|
| 379 |
+
"learning_rate": 1.168241965973535e-05,
|
| 380 |
+
"loss": 0.0269,
|
| 381 |
+
"step": 4400
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 8.51,
|
| 385 |
+
"grad_norm": 0.001653732149861753,
|
| 386 |
+
"learning_rate": 1.1493383742911156e-05,
|
| 387 |
+
"loss": 0.0208,
|
| 388 |
+
"step": 4500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 8.7,
|
| 392 |
+
"grad_norm": 0.002103479579091072,
|
| 393 |
+
"learning_rate": 1.1304347826086957e-05,
|
| 394 |
+
"loss": 0.0263,
|
| 395 |
+
"step": 4600
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 8.88,
|
| 399 |
+
"grad_norm": 0.001875244197435677,
|
| 400 |
+
"learning_rate": 1.111531190926276e-05,
|
| 401 |
+
"loss": 0.0086,
|
| 402 |
+
"step": 4700
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 9.0,
|
| 406 |
+
"eval_loss": 0.3807547688484192,
|
| 407 |
+
"eval_runtime": 4.2581,
|
| 408 |
+
"eval_samples_per_second": 91.355,
|
| 409 |
+
"eval_steps_per_second": 6.576,
|
| 410 |
+
"step": 4761
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 9.07,
|
| 414 |
+
"grad_norm": 0.0033042938448488712,
|
| 415 |
+
"learning_rate": 1.0926275992438564e-05,
|
| 416 |
+
"loss": 0.0361,
|
| 417 |
+
"step": 4800
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 9.26,
|
| 421 |
+
"grad_norm": 0.0017488696612417698,
|
| 422 |
+
"learning_rate": 1.0737240075614367e-05,
|
| 423 |
+
"loss": 0.0194,
|
| 424 |
+
"step": 4900
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 9.45,
|
| 428 |
+
"grad_norm": 0.50156569480896,
|
| 429 |
+
"learning_rate": 1.0548204158790173e-05,
|
| 430 |
+
"loss": 0.0338,
|
| 431 |
+
"step": 5000
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 9.64,
|
| 435 |
+
"grad_norm": 0.0015125612262636423,
|
| 436 |
+
"learning_rate": 1.0359168241965973e-05,
|
| 437 |
+
"loss": 0.0195,
|
| 438 |
+
"step": 5100
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 9.83,
|
| 442 |
+
"grad_norm": 0.0021789916791021824,
|
| 443 |
+
"learning_rate": 1.0170132325141778e-05,
|
| 444 |
+
"loss": 0.0114,
|
| 445 |
+
"step": 5200
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 10.0,
|
| 449 |
+
"eval_loss": 0.5712518095970154,
|
| 450 |
+
"eval_runtime": 4.2506,
|
| 451 |
+
"eval_samples_per_second": 91.516,
|
| 452 |
+
"eval_steps_per_second": 6.587,
|
| 453 |
+
"step": 5290
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 10.02,
|
| 457 |
+
"grad_norm": 0.000900177750736475,
|
| 458 |
+
"learning_rate": 9.981096408317581e-06,
|
| 459 |
+
"loss": 0.0229,
|
| 460 |
+
"step": 5300
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 10.21,
|
| 464 |
+
"grad_norm": 0.004578994121402502,
|
| 465 |
+
"learning_rate": 9.792060491493385e-06,
|
| 466 |
+
"loss": 0.0226,
|
| 467 |
+
"step": 5400
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 10.4,
|
| 471 |
+
"grad_norm": 0.0009541900944896042,
|
| 472 |
+
"learning_rate": 9.603024574669188e-06,
|
| 473 |
+
"loss": 0.0197,
|
| 474 |
+
"step": 5500
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 10.59,
|
| 478 |
+
"grad_norm": 0.000505660951603204,
|
| 479 |
+
"learning_rate": 9.413988657844991e-06,
|
| 480 |
+
"loss": 0.016,
|
| 481 |
+
"step": 5600
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 10.78,
|
| 485 |
+
"grad_norm": 0.001706490176729858,
|
| 486 |
+
"learning_rate": 9.224952741020795e-06,
|
| 487 |
+
"loss": 0.0188,
|
| 488 |
+
"step": 5700
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 10.96,
|
| 492 |
+
"grad_norm": 0.015518520027399063,
|
| 493 |
+
"learning_rate": 9.035916824196598e-06,
|
| 494 |
+
"loss": 0.0214,
|
| 495 |
+
"step": 5800
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 11.0,
|
| 499 |
+
"eval_loss": 0.48657986521720886,
|
| 500 |
+
"eval_runtime": 4.2449,
|
| 501 |
+
"eval_samples_per_second": 91.64,
|
| 502 |
+
"eval_steps_per_second": 6.596,
|
| 503 |
+
"step": 5819
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 11.15,
|
| 507 |
+
"grad_norm": 0.006599494256079197,
|
| 508 |
+
"learning_rate": 8.846880907372402e-06,
|
| 509 |
+
"loss": 0.0188,
|
| 510 |
+
"step": 5900
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 11.34,
|
| 514 |
+
"grad_norm": 0.002344762207940221,
|
| 515 |
+
"learning_rate": 8.657844990548205e-06,
|
| 516 |
+
"loss": 0.0374,
|
| 517 |
+
"step": 6000
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 11.53,
|
| 521 |
+
"grad_norm": 0.0021403091959655285,
|
| 522 |
+
"learning_rate": 8.468809073724009e-06,
|
| 523 |
+
"loss": 0.0331,
|
| 524 |
+
"step": 6100
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 11.72,
|
| 528 |
+
"grad_norm": 0.638529360294342,
|
| 529 |
+
"learning_rate": 8.279773156899812e-06,
|
| 530 |
+
"loss": 0.0058,
|
| 531 |
+
"step": 6200
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 11.91,
|
| 535 |
+
"grad_norm": 0.0011522163404151797,
|
| 536 |
+
"learning_rate": 8.090737240075615e-06,
|
| 537 |
+
"loss": 0.0169,
|
| 538 |
+
"step": 6300
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 12.0,
|
| 542 |
+
"eval_loss": 0.514002799987793,
|
| 543 |
+
"eval_runtime": 4.2661,
|
| 544 |
+
"eval_samples_per_second": 91.185,
|
| 545 |
+
"eval_steps_per_second": 6.563,
|
| 546 |
+
"step": 6348
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 12.1,
|
| 550 |
+
"grad_norm": 0.0009117820300161839,
|
| 551 |
+
"learning_rate": 7.901701323251419e-06,
|
| 552 |
+
"loss": 0.0139,
|
| 553 |
+
"step": 6400
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 12.29,
|
| 557 |
+
"grad_norm": 0.05164305865764618,
|
| 558 |
+
"learning_rate": 7.712665406427222e-06,
|
| 559 |
+
"loss": 0.0152,
|
| 560 |
+
"step": 6500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 12.48,
|
| 564 |
+
"grad_norm": 22.255159378051758,
|
| 565 |
+
"learning_rate": 7.523629489603025e-06,
|
| 566 |
+
"loss": 0.018,
|
| 567 |
+
"step": 6600
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 12.67,
|
| 571 |
+
"grad_norm": 0.0009817625395953655,
|
| 572 |
+
"learning_rate": 7.334593572778829e-06,
|
| 573 |
+
"loss": 0.0092,
|
| 574 |
+
"step": 6700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 12.85,
|
| 578 |
+
"grad_norm": 0.00901391077786684,
|
| 579 |
+
"learning_rate": 7.1455576559546324e-06,
|
| 580 |
+
"loss": 0.0247,
|
| 581 |
+
"step": 6800
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 13.0,
|
| 585 |
+
"eval_loss": 0.40500932931900024,
|
| 586 |
+
"eval_runtime": 4.2563,
|
| 587 |
+
"eval_samples_per_second": 91.395,
|
| 588 |
+
"eval_steps_per_second": 6.579,
|
| 589 |
+
"step": 6877
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 13.04,
|
| 593 |
+
"grad_norm": 0.002569295698776841,
|
| 594 |
+
"learning_rate": 6.956521739130435e-06,
|
| 595 |
+
"loss": 0.0228,
|
| 596 |
+
"step": 6900
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 13.23,
|
| 600 |
+
"grad_norm": 0.0035759450402110815,
|
| 601 |
+
"learning_rate": 6.767485822306239e-06,
|
| 602 |
+
"loss": 0.0166,
|
| 603 |
+
"step": 7000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 13.42,
|
| 607 |
+
"grad_norm": 0.0027633486315608025,
|
| 608 |
+
"learning_rate": 6.578449905482042e-06,
|
| 609 |
+
"loss": 0.0185,
|
| 610 |
+
"step": 7100
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 13.61,
|
| 614 |
+
"grad_norm": 0.0013614584458991885,
|
| 615 |
+
"learning_rate": 6.389413988657845e-06,
|
| 616 |
+
"loss": 0.0088,
|
| 617 |
+
"step": 7200
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 13.8,
|
| 621 |
+
"grad_norm": 0.002298623789101839,
|
| 622 |
+
"learning_rate": 6.2003780718336495e-06,
|
| 623 |
+
"loss": 0.0152,
|
| 624 |
+
"step": 7300
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 13.99,
|
| 628 |
+
"grad_norm": 0.0007343721226789057,
|
| 629 |
+
"learning_rate": 6.011342155009452e-06,
|
| 630 |
+
"loss": 0.0082,
|
| 631 |
+
"step": 7400
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 14.0,
|
| 635 |
+
"eval_loss": 0.48581066727638245,
|
| 636 |
+
"eval_runtime": 4.2624,
|
| 637 |
+
"eval_samples_per_second": 91.264,
|
| 638 |
+
"eval_steps_per_second": 6.569,
|
| 639 |
+
"step": 7406
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 14.18,
|
| 643 |
+
"grad_norm": 0.009072878398001194,
|
| 644 |
+
"learning_rate": 5.8223062381852555e-06,
|
| 645 |
+
"loss": 0.0116,
|
| 646 |
+
"step": 7500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 14.37,
|
| 650 |
+
"grad_norm": 0.0018677357584238052,
|
| 651 |
+
"learning_rate": 5.63327032136106e-06,
|
| 652 |
+
"loss": 0.0083,
|
| 653 |
+
"step": 7600
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 14.56,
|
| 657 |
+
"grad_norm": 0.003356189001351595,
|
| 658 |
+
"learning_rate": 5.444234404536862e-06,
|
| 659 |
+
"loss": 0.0148,
|
| 660 |
+
"step": 7700
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 14.74,
|
| 664 |
+
"grad_norm": 0.011417761445045471,
|
| 665 |
+
"learning_rate": 5.255198487712666e-06,
|
| 666 |
+
"loss": 0.014,
|
| 667 |
+
"step": 7800
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 14.93,
|
| 671 |
+
"grad_norm": 0.0009720730595290661,
|
| 672 |
+
"learning_rate": 5.066162570888468e-06,
|
| 673 |
+
"loss": 0.0058,
|
| 674 |
+
"step": 7900
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 15.0,
|
| 678 |
+
"eval_loss": 0.5540414452552795,
|
| 679 |
+
"eval_runtime": 4.242,
|
| 680 |
+
"eval_samples_per_second": 91.701,
|
| 681 |
+
"eval_steps_per_second": 6.601,
|
| 682 |
+
"step": 7935
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 15.12,
|
| 686 |
+
"grad_norm": 0.0007459772750735283,
|
| 687 |
+
"learning_rate": 4.877126654064273e-06,
|
| 688 |
+
"loss": 0.0061,
|
| 689 |
+
"step": 8000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 15.31,
|
| 693 |
+
"grad_norm": 0.0017996943788602948,
|
| 694 |
+
"learning_rate": 4.688090737240076e-06,
|
| 695 |
+
"loss": 0.0133,
|
| 696 |
+
"step": 8100
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 15.5,
|
| 700 |
+
"grad_norm": 0.0010721386643126607,
|
| 701 |
+
"learning_rate": 4.4990548204158795e-06,
|
| 702 |
+
"loss": 0.0075,
|
| 703 |
+
"step": 8200
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 15.69,
|
| 707 |
+
"grad_norm": 0.0009099834132939577,
|
| 708 |
+
"learning_rate": 4.310018903591683e-06,
|
| 709 |
+
"loss": 0.0026,
|
| 710 |
+
"step": 8300
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 15.88,
|
| 714 |
+
"grad_norm": 0.00020547083113342524,
|
| 715 |
+
"learning_rate": 4.120982986767486e-06,
|
| 716 |
+
"loss": 0.0,
|
| 717 |
+
"step": 8400
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 16.0,
|
| 721 |
+
"eval_loss": 0.5253579020500183,
|
| 722 |
+
"eval_runtime": 4.2546,
|
| 723 |
+
"eval_samples_per_second": 91.43,
|
| 724 |
+
"eval_steps_per_second": 6.581,
|
| 725 |
+
"step": 8464
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 16.07,
|
| 729 |
+
"grad_norm": 0.0004376711731310934,
|
| 730 |
+
"learning_rate": 3.93194706994329e-06,
|
| 731 |
+
"loss": 0.0073,
|
| 732 |
+
"step": 8500
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 16.26,
|
| 736 |
+
"grad_norm": 0.0015534511767327785,
|
| 737 |
+
"learning_rate": 3.742911153119093e-06,
|
| 738 |
+
"loss": 0.0053,
|
| 739 |
+
"step": 8600
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 16.45,
|
| 743 |
+
"grad_norm": 0.0005960518028587103,
|
| 744 |
+
"learning_rate": 3.5538752362948966e-06,
|
| 745 |
+
"loss": 0.0,
|
| 746 |
+
"step": 8700
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 16.64,
|
| 750 |
+
"grad_norm": 0.00040510049439035356,
|
| 751 |
+
"learning_rate": 3.3648393194706996e-06,
|
| 752 |
+
"loss": 0.0,
|
| 753 |
+
"step": 8800
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 16.82,
|
| 757 |
+
"grad_norm": 0.0001634809304960072,
|
| 758 |
+
"learning_rate": 3.175803402646503e-06,
|
| 759 |
+
"loss": 0.0,
|
| 760 |
+
"step": 8900
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 17.0,
|
| 764 |
+
"eval_loss": 0.574542760848999,
|
| 765 |
+
"eval_runtime": 4.2742,
|
| 766 |
+
"eval_samples_per_second": 91.011,
|
| 767 |
+
"eval_steps_per_second": 6.551,
|
| 768 |
+
"step": 8993
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 17.01,
|
| 772 |
+
"grad_norm": 0.0002437944058328867,
|
| 773 |
+
"learning_rate": 2.986767485822307e-06,
|
| 774 |
+
"loss": 0.0104,
|
| 775 |
+
"step": 9000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 17.2,
|
| 779 |
+
"grad_norm": 0.00035620402195490897,
|
| 780 |
+
"learning_rate": 2.79773156899811e-06,
|
| 781 |
+
"loss": 0.0001,
|
| 782 |
+
"step": 9100
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 17.39,
|
| 786 |
+
"grad_norm": 0.00021548435324802995,
|
| 787 |
+
"learning_rate": 2.6086956521739132e-06,
|
| 788 |
+
"loss": 0.0037,
|
| 789 |
+
"step": 9200
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 17.58,
|
| 793 |
+
"grad_norm": 0.00014057717635296285,
|
| 794 |
+
"learning_rate": 2.4196597353497167e-06,
|
| 795 |
+
"loss": 0.0,
|
| 796 |
+
"step": 9300
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 17.77,
|
| 800 |
+
"grad_norm": 0.0010042821522802114,
|
| 801 |
+
"learning_rate": 2.23062381852552e-06,
|
| 802 |
+
"loss": 0.0109,
|
| 803 |
+
"step": 9400
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 17.96,
|
| 807 |
+
"grad_norm": 0.00016069511184468865,
|
| 808 |
+
"learning_rate": 2.0415879017013235e-06,
|
| 809 |
+
"loss": 0.0,
|
| 810 |
+
"step": 9500
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 18.0,
|
| 814 |
+
"eval_loss": 0.5317645072937012,
|
| 815 |
+
"eval_runtime": 4.2532,
|
| 816 |
+
"eval_samples_per_second": 91.461,
|
| 817 |
+
"eval_steps_per_second": 6.583,
|
| 818 |
+
"step": 9522
|
| 819 |
+
}
|
| 820 |
+
],
|
| 821 |
+
"logging_steps": 100,
|
| 822 |
+
"max_steps": 10580,
|
| 823 |
+
"num_input_tokens_seen": 0,
|
| 824 |
+
"num_train_epochs": 20,
|
| 825 |
+
"save_steps": 500,
|
| 826 |
+
"total_flos": 9.72342785672254e+16,
|
| 827 |
+
"train_batch_size": 14,
|
| 828 |
+
"trial_name": null,
|
| 829 |
+
"trial_params": null
|
| 830 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e38d84d4a8a34ed1c3d4c1693843cfd19f859389604ec942e9d423aad70c82fc
|
| 3 |
+
size 4856
|