Upload 7 files
Browse files- config.json +47 -0
- model.safetensors +3 -0
- optimizer-002.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +825 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "answerdotai/ModernBERT-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ModernBERTForMultilabel"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 50281,
|
| 9 |
+
"classifier_activation": "gelu",
|
| 10 |
+
"classifier_bias": false,
|
| 11 |
+
"classifier_dropout": 0.0,
|
| 12 |
+
"classifier_pooling": "mean",
|
| 13 |
+
"cls_token_id": 50281,
|
| 14 |
+
"decoder_bias": true,
|
| 15 |
+
"deterministic_flash_attn": false,
|
| 16 |
+
"embedding_dropout": 0.0,
|
| 17 |
+
"eos_token_id": 50282,
|
| 18 |
+
"global_attn_every_n_layers": 3,
|
| 19 |
+
"global_rope_theta": 160000.0,
|
| 20 |
+
"gradient_checkpointing": false,
|
| 21 |
+
"hidden_activation": "gelu",
|
| 22 |
+
"hidden_size": 1024,
|
| 23 |
+
"initializer_cutoff_factor": 2.0,
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"intermediate_size": 2624,
|
| 26 |
+
"layer_norm_eps": 1e-05,
|
| 27 |
+
"local_attention": 128,
|
| 28 |
+
"local_rope_theta": 10000.0,
|
| 29 |
+
"max_position_embeddings": 8192,
|
| 30 |
+
"mlp_bias": false,
|
| 31 |
+
"mlp_dropout": 0.0,
|
| 32 |
+
"model_type": "modernbert",
|
| 33 |
+
"norm_bias": false,
|
| 34 |
+
"norm_eps": 1e-05,
|
| 35 |
+
"num_attention_heads": 16,
|
| 36 |
+
"num_hidden_layers": 28,
|
| 37 |
+
"pad_token_id": 50283,
|
| 38 |
+
"position_embedding_type": "absolute",
|
| 39 |
+
"reference_compile": false,
|
| 40 |
+
"repad_logits_with_grad": false,
|
| 41 |
+
"sep_token_id": 50282,
|
| 42 |
+
"sparse_pred_ignore_index": -100,
|
| 43 |
+
"sparse_prediction": false,
|
| 44 |
+
"torch_dtype": "float32",
|
| 45 |
+
"transformers_version": "4.48.3",
|
| 46 |
+
"vocab_size": 50368
|
| 47 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f8b6e22d13a2e337eb0d9d9260318f9b9e66894ed7307f827fda9bbe3cb0e86
|
| 3 |
+
size 1583364804
|
optimizer-002.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8cef05dab1da46a3ddffaba28f17d27107e236831151418b7e5a0c9da94a2eb
|
| 3 |
+
size 3166837754
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80bfdc58b7f847bcb148b9d44718eca85d190fadb2e5b8f03ea078e20bc68b54
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81d60e0935192c9d68d019378097554cfd188c9945c392a8cf7a3bc06d8ef236
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,825 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8000965798413886,
|
| 3 |
+
"best_model_checkpoint": "/content/drive/MyDrive/Projects/ModernBertTuning/2025-02-19_17:10:03/checkpoint-3300",
|
| 4 |
+
"epoch": 5.809859154929578,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 3300,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.176056338028169,
|
| 13 |
+
"grad_norm": 5.5684285163879395,
|
| 14 |
+
"learning_rate": 8.802816901408451e-06,
|
| 15 |
+
"loss": 0.9029,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.176056338028169,
|
| 20 |
+
"eval_action_accuracy": 0.7632042253521126,
|
| 21 |
+
"eval_action_f1": 0.3275,
|
| 22 |
+
"eval_action_precision": 0.6787564766839378,
|
| 23 |
+
"eval_action_recall": 0.2158154859967051,
|
| 24 |
+
"eval_combined_f1": 0.39009832159236646,
|
| 25 |
+
"eval_loss": 1.5983511209487915,
|
| 26 |
+
"eval_runtime": 53.3406,
|
| 27 |
+
"eval_samples_per_second": 42.594,
|
| 28 |
+
"eval_severity_accuracy": 0.4942781690140845,
|
| 29 |
+
"eval_severity_f1": 0.4526966431847329,
|
| 30 |
+
"eval_severity_precision": 0.5190575556900403,
|
| 31 |
+
"eval_severity_recall": 0.4942781690140845,
|
| 32 |
+
"eval_steps_per_second": 2.662,
|
| 33 |
+
"step": 100
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.352112676056338,
|
| 37 |
+
"grad_norm": 10.437986373901367,
|
| 38 |
+
"learning_rate": 1.7605633802816902e-05,
|
| 39 |
+
"loss": 0.6875,
|
| 40 |
+
"step": 200
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.352112676056338,
|
| 44 |
+
"eval_action_accuracy": 0.8393485915492958,
|
| 45 |
+
"eval_action_f1": 0.7253574115876599,
|
| 46 |
+
"eval_action_precision": 0.667590027700831,
|
| 47 |
+
"eval_action_recall": 0.7940691927512356,
|
| 48 |
+
"eval_combined_f1": 0.6889564202307465,
|
| 49 |
+
"eval_loss": 1.1708728075027466,
|
| 50 |
+
"eval_runtime": 53.402,
|
| 51 |
+
"eval_samples_per_second": 42.545,
|
| 52 |
+
"eval_severity_accuracy": 0.6615316901408451,
|
| 53 |
+
"eval_severity_f1": 0.6525554288738333,
|
| 54 |
+
"eval_severity_precision": 0.6774842234837531,
|
| 55 |
+
"eval_severity_recall": 0.6615316901408451,
|
| 56 |
+
"eval_steps_per_second": 2.659,
|
| 57 |
+
"step": 200
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.528169014084507,
|
| 61 |
+
"grad_norm": 8.755282402038574,
|
| 62 |
+
"learning_rate": 2.640845070422535e-05,
|
| 63 |
+
"loss": 0.5914,
|
| 64 |
+
"step": 300
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.528169014084507,
|
| 68 |
+
"eval_action_accuracy": 0.8653169014084507,
|
| 69 |
+
"eval_action_f1": 0.7688821752265861,
|
| 70 |
+
"eval_action_precision": 0.7099023709902371,
|
| 71 |
+
"eval_action_recall": 0.8385502471169687,
|
| 72 |
+
"eval_combined_f1": 0.7012112761418396,
|
| 73 |
+
"eval_loss": 1.1276360750198364,
|
| 74 |
+
"eval_runtime": 53.4046,
|
| 75 |
+
"eval_samples_per_second": 42.543,
|
| 76 |
+
"eval_severity_accuracy": 0.6307218309859155,
|
| 77 |
+
"eval_severity_f1": 0.6335403770570933,
|
| 78 |
+
"eval_severity_precision": 0.6726339478831171,
|
| 79 |
+
"eval_severity_recall": 0.6307218309859155,
|
| 80 |
+
"eval_steps_per_second": 2.659,
|
| 81 |
+
"step": 300
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.704225352112676,
|
| 85 |
+
"grad_norm": 4.745133399963379,
|
| 86 |
+
"learning_rate": 3.5211267605633805e-05,
|
| 87 |
+
"loss": 0.5167,
|
| 88 |
+
"step": 400
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.704225352112676,
|
| 92 |
+
"eval_action_accuracy": 0.8820422535211268,
|
| 93 |
+
"eval_action_f1": 0.750465549348231,
|
| 94 |
+
"eval_action_precision": 0.8629550321199143,
|
| 95 |
+
"eval_action_recall": 0.6639209225700164,
|
| 96 |
+
"eval_combined_f1": 0.7324396418875856,
|
| 97 |
+
"eval_loss": 1.0260754823684692,
|
| 98 |
+
"eval_runtime": 53.3961,
|
| 99 |
+
"eval_samples_per_second": 42.55,
|
| 100 |
+
"eval_severity_accuracy": 0.715669014084507,
|
| 101 |
+
"eval_severity_f1": 0.7144137344269402,
|
| 102 |
+
"eval_severity_precision": 0.717618479930179,
|
| 103 |
+
"eval_severity_recall": 0.715669014084507,
|
| 104 |
+
"eval_steps_per_second": 2.659,
|
| 105 |
+
"step": 400
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.8802816901408451,
|
| 109 |
+
"grad_norm": 4.491367816925049,
|
| 110 |
+
"learning_rate": 4.4014084507042256e-05,
|
| 111 |
+
"loss": 0.5228,
|
| 112 |
+
"step": 500
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.8802816901408451,
|
| 116 |
+
"eval_action_accuracy": 0.8811619718309859,
|
| 117 |
+
"eval_action_f1": 0.7652173913043478,
|
| 118 |
+
"eval_action_precision": 0.8103130755064457,
|
| 119 |
+
"eval_action_recall": 0.7248764415156508,
|
| 120 |
+
"eval_combined_f1": 0.7303003562324298,
|
| 121 |
+
"eval_loss": 1.0010697841644287,
|
| 122 |
+
"eval_runtime": 53.3858,
|
| 123 |
+
"eval_samples_per_second": 42.558,
|
| 124 |
+
"eval_severity_accuracy": 0.6941021126760564,
|
| 125 |
+
"eval_severity_f1": 0.6953833211605118,
|
| 126 |
+
"eval_severity_precision": 0.7027972731452775,
|
| 127 |
+
"eval_severity_recall": 0.6941021126760564,
|
| 128 |
+
"eval_steps_per_second": 2.66,
|
| 129 |
+
"step": 500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.056338028169014,
|
| 133 |
+
"grad_norm": 7.762014865875244,
|
| 134 |
+
"learning_rate": 4.968701095461659e-05,
|
| 135 |
+
"loss": 0.4707,
|
| 136 |
+
"step": 600
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.056338028169014,
|
| 140 |
+
"eval_action_accuracy": 0.8816021126760564,
|
| 141 |
+
"eval_action_f1": 0.7497674418604651,
|
| 142 |
+
"eval_action_precision": 0.8611111111111112,
|
| 143 |
+
"eval_action_recall": 0.6639209225700164,
|
| 144 |
+
"eval_combined_f1": 0.7288028162963264,
|
| 145 |
+
"eval_loss": 1.05709969997406,
|
| 146 |
+
"eval_runtime": 53.3852,
|
| 147 |
+
"eval_samples_per_second": 42.559,
|
| 148 |
+
"eval_severity_accuracy": 0.710387323943662,
|
| 149 |
+
"eval_severity_f1": 0.7078381907321877,
|
| 150 |
+
"eval_severity_precision": 0.7185583041734807,
|
| 151 |
+
"eval_severity_recall": 0.710387323943662,
|
| 152 |
+
"eval_steps_per_second": 2.66,
|
| 153 |
+
"step": 600
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.232394366197183,
|
| 157 |
+
"grad_norm": 9.271745681762695,
|
| 158 |
+
"learning_rate": 4.870892018779343e-05,
|
| 159 |
+
"loss": 0.4103,
|
| 160 |
+
"step": 700
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.232394366197183,
|
| 164 |
+
"eval_action_accuracy": 0.8503521126760564,
|
| 165 |
+
"eval_action_f1": 0.621380846325167,
|
| 166 |
+
"eval_action_precision": 0.9587628865979382,
|
| 167 |
+
"eval_action_recall": 0.4596375617792422,
|
| 168 |
+
"eval_combined_f1": 0.6340972974514396,
|
| 169 |
+
"eval_loss": 1.1522221565246582,
|
| 170 |
+
"eval_runtime": 53.3981,
|
| 171 |
+
"eval_samples_per_second": 42.548,
|
| 172 |
+
"eval_severity_accuracy": 0.6553697183098591,
|
| 173 |
+
"eval_severity_f1": 0.6468137485777123,
|
| 174 |
+
"eval_severity_precision": 0.7014511458318834,
|
| 175 |
+
"eval_severity_recall": 0.6553697183098591,
|
| 176 |
+
"eval_steps_per_second": 2.659,
|
| 177 |
+
"step": 700
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.408450704225352,
|
| 181 |
+
"grad_norm": 7.816728115081787,
|
| 182 |
+
"learning_rate": 4.773082942097027e-05,
|
| 183 |
+
"loss": 0.4,
|
| 184 |
+
"step": 800
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 1.408450704225352,
|
| 188 |
+
"eval_action_accuracy": 0.8820422535211268,
|
| 189 |
+
"eval_action_f1": 0.7922480620155039,
|
| 190 |
+
"eval_action_precision": 0.7481698389458272,
|
| 191 |
+
"eval_action_recall": 0.841845140032949,
|
| 192 |
+
"eval_combined_f1": 0.7584899763574804,
|
| 193 |
+
"eval_loss": 0.9812976717948914,
|
| 194 |
+
"eval_runtime": 53.3978,
|
| 195 |
+
"eval_samples_per_second": 42.549,
|
| 196 |
+
"eval_severity_accuracy": 0.7235915492957746,
|
| 197 |
+
"eval_severity_f1": 0.724731890699457,
|
| 198 |
+
"eval_severity_precision": 0.7317594548978289,
|
| 199 |
+
"eval_severity_recall": 0.7235915492957746,
|
| 200 |
+
"eval_steps_per_second": 2.659,
|
| 201 |
+
"step": 800
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.584507042253521,
|
| 205 |
+
"grad_norm": 7.098080158233643,
|
| 206 |
+
"learning_rate": 4.675273865414711e-05,
|
| 207 |
+
"loss": 0.4179,
|
| 208 |
+
"step": 900
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.584507042253521,
|
| 212 |
+
"eval_action_accuracy": 0.894806338028169,
|
| 213 |
+
"eval_action_f1": 0.7879325643300799,
|
| 214 |
+
"eval_action_precision": 0.8538461538461538,
|
| 215 |
+
"eval_action_recall": 0.7314662273476112,
|
| 216 |
+
"eval_combined_f1": 0.7270919018906684,
|
| 217 |
+
"eval_loss": 1.177077054977417,
|
| 218 |
+
"eval_runtime": 53.4076,
|
| 219 |
+
"eval_samples_per_second": 42.541,
|
| 220 |
+
"eval_severity_accuracy": 0.6822183098591549,
|
| 221 |
+
"eval_severity_f1": 0.666251239451257,
|
| 222 |
+
"eval_severity_precision": 0.7051808476907062,
|
| 223 |
+
"eval_severity_recall": 0.6822183098591549,
|
| 224 |
+
"eval_steps_per_second": 2.659,
|
| 225 |
+
"step": 900
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 1.76056338028169,
|
| 229 |
+
"grad_norm": 3.6362385749816895,
|
| 230 |
+
"learning_rate": 4.577464788732395e-05,
|
| 231 |
+
"loss": 0.381,
|
| 232 |
+
"step": 1000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 1.76056338028169,
|
| 236 |
+
"eval_action_accuracy": 0.8934859154929577,
|
| 237 |
+
"eval_action_f1": 0.8,
|
| 238 |
+
"eval_action_precision": 0.802653399668325,
|
| 239 |
+
"eval_action_recall": 0.7973640856672158,
|
| 240 |
+
"eval_combined_f1": 0.7276179625363967,
|
| 241 |
+
"eval_loss": 1.0336343050003052,
|
| 242 |
+
"eval_runtime": 53.3973,
|
| 243 |
+
"eval_samples_per_second": 42.549,
|
| 244 |
+
"eval_severity_accuracy": 0.6597711267605634,
|
| 245 |
+
"eval_severity_f1": 0.6552359250727933,
|
| 246 |
+
"eval_severity_precision": 0.690194073616287,
|
| 247 |
+
"eval_severity_recall": 0.6597711267605634,
|
| 248 |
+
"eval_steps_per_second": 2.659,
|
| 249 |
+
"step": 1000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.936619718309859,
|
| 253 |
+
"grad_norm": 4.175089359283447,
|
| 254 |
+
"learning_rate": 4.479655712050079e-05,
|
| 255 |
+
"loss": 0.384,
|
| 256 |
+
"step": 1100
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.936619718309859,
|
| 260 |
+
"eval_action_accuracy": 0.8899647887323944,
|
| 261 |
+
"eval_action_f1": 0.7987117552334944,
|
| 262 |
+
"eval_action_precision": 0.7811023622047244,
|
| 263 |
+
"eval_action_recall": 0.8171334431630972,
|
| 264 |
+
"eval_combined_f1": 0.7613638710145674,
|
| 265 |
+
"eval_loss": 0.9859868288040161,
|
| 266 |
+
"eval_runtime": 53.3975,
|
| 267 |
+
"eval_samples_per_second": 42.549,
|
| 268 |
+
"eval_severity_accuracy": 0.7240316901408451,
|
| 269 |
+
"eval_severity_f1": 0.7240159867956404,
|
| 270 |
+
"eval_severity_precision": 0.7333723870533466,
|
| 271 |
+
"eval_severity_recall": 0.7240316901408451,
|
| 272 |
+
"eval_steps_per_second": 2.659,
|
| 273 |
+
"step": 1100
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 2.112676056338028,
|
| 277 |
+
"grad_norm": 5.42644739151001,
|
| 278 |
+
"learning_rate": 4.3818466353677626e-05,
|
| 279 |
+
"loss": 0.287,
|
| 280 |
+
"step": 1200
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 2.112676056338028,
|
| 284 |
+
"eval_action_accuracy": 0.8798415492957746,
|
| 285 |
+
"eval_action_f1": 0.7272727272727273,
|
| 286 |
+
"eval_action_precision": 0.9238578680203046,
|
| 287 |
+
"eval_action_recall": 0.5996705107084019,
|
| 288 |
+
"eval_combined_f1": 0.7194719198906718,
|
| 289 |
+
"eval_loss": 1.2687214612960815,
|
| 290 |
+
"eval_runtime": 53.3984,
|
| 291 |
+
"eval_samples_per_second": 42.548,
|
| 292 |
+
"eval_severity_accuracy": 0.7130281690140845,
|
| 293 |
+
"eval_severity_f1": 0.7116711125086165,
|
| 294 |
+
"eval_severity_precision": 0.7299090747607111,
|
| 295 |
+
"eval_severity_recall": 0.7130281690140845,
|
| 296 |
+
"eval_steps_per_second": 2.659,
|
| 297 |
+
"step": 1200
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 2.288732394366197,
|
| 301 |
+
"grad_norm": 4.9831342697143555,
|
| 302 |
+
"learning_rate": 4.284037558685447e-05,
|
| 303 |
+
"loss": 0.234,
|
| 304 |
+
"step": 1300
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 2.288732394366197,
|
| 308 |
+
"eval_action_accuracy": 0.8758802816901409,
|
| 309 |
+
"eval_action_f1": 0.7904903417533432,
|
| 310 |
+
"eval_action_precision": 0.7198917456021651,
|
| 311 |
+
"eval_action_recall": 0.8764415156507414,
|
| 312 |
+
"eval_combined_f1": 0.7408713797328776,
|
| 313 |
+
"eval_loss": 1.2081983089447021,
|
| 314 |
+
"eval_runtime": 53.3856,
|
| 315 |
+
"eval_samples_per_second": 42.558,
|
| 316 |
+
"eval_severity_accuracy": 0.6954225352112676,
|
| 317 |
+
"eval_severity_f1": 0.691252417712412,
|
| 318 |
+
"eval_severity_precision": 0.710940220005423,
|
| 319 |
+
"eval_severity_recall": 0.6954225352112676,
|
| 320 |
+
"eval_steps_per_second": 2.66,
|
| 321 |
+
"step": 1300
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.464788732394366,
|
| 325 |
+
"grad_norm": 8.448158264160156,
|
| 326 |
+
"learning_rate": 4.18622848200313e-05,
|
| 327 |
+
"loss": 0.2284,
|
| 328 |
+
"step": 1400
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.464788732394366,
|
| 332 |
+
"eval_action_accuracy": 0.9097711267605634,
|
| 333 |
+
"eval_action_f1": 0.820333041191937,
|
| 334 |
+
"eval_action_precision": 0.8764044943820225,
|
| 335 |
+
"eval_action_recall": 0.771004942339374,
|
| 336 |
+
"eval_combined_f1": 0.7709403023211041,
|
| 337 |
+
"eval_loss": 1.1236168146133423,
|
| 338 |
+
"eval_runtime": 53.3827,
|
| 339 |
+
"eval_samples_per_second": 42.561,
|
| 340 |
+
"eval_severity_accuracy": 0.7205105633802817,
|
| 341 |
+
"eval_severity_f1": 0.7215475634502713,
|
| 342 |
+
"eval_severity_precision": 0.7388563066943578,
|
| 343 |
+
"eval_severity_recall": 0.7205105633802817,
|
| 344 |
+
"eval_steps_per_second": 2.66,
|
| 345 |
+
"step": 1400
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.640845070422535,
|
| 349 |
+
"grad_norm": 3.273317575454712,
|
| 350 |
+
"learning_rate": 4.0884194053208144e-05,
|
| 351 |
+
"loss": 0.2147,
|
| 352 |
+
"step": 1500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.640845070422535,
|
| 356 |
+
"eval_action_accuracy": 0.9066901408450704,
|
| 357 |
+
"eval_action_f1": 0.8233333333333334,
|
| 358 |
+
"eval_action_precision": 0.8330522765598651,
|
| 359 |
+
"eval_action_recall": 0.8138385502471169,
|
| 360 |
+
"eval_combined_f1": 0.7802089055740911,
|
| 361 |
+
"eval_loss": 1.08566153049469,
|
| 362 |
+
"eval_runtime": 53.3876,
|
| 363 |
+
"eval_samples_per_second": 42.557,
|
| 364 |
+
"eval_severity_accuracy": 0.7394366197183099,
|
| 365 |
+
"eval_severity_f1": 0.7370844778148489,
|
| 366 |
+
"eval_severity_precision": 0.7443892609781334,
|
| 367 |
+
"eval_severity_recall": 0.7394366197183099,
|
| 368 |
+
"eval_steps_per_second": 2.66,
|
| 369 |
+
"step": 1500
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.816901408450704,
|
| 373 |
+
"grad_norm": 4.577557563781738,
|
| 374 |
+
"learning_rate": 3.990610328638498e-05,
|
| 375 |
+
"loss": 0.2258,
|
| 376 |
+
"step": 1600
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 2.816901408450704,
|
| 380 |
+
"eval_action_accuracy": 0.8939260563380281,
|
| 381 |
+
"eval_action_f1": 0.7872903795233892,
|
| 382 |
+
"eval_action_precision": 0.8479087452471483,
|
| 383 |
+
"eval_action_recall": 0.7347611202635914,
|
| 384 |
+
"eval_combined_f1": 0.7659288423194548,
|
| 385 |
+
"eval_loss": 1.0919924974441528,
|
| 386 |
+
"eval_runtime": 53.4132,
|
| 387 |
+
"eval_samples_per_second": 42.536,
|
| 388 |
+
"eval_severity_accuracy": 0.7442781690140845,
|
| 389 |
+
"eval_severity_f1": 0.7445673051155204,
|
| 390 |
+
"eval_severity_precision": 0.7537282342243937,
|
| 391 |
+
"eval_severity_recall": 0.7442781690140845,
|
| 392 |
+
"eval_steps_per_second": 2.659,
|
| 393 |
+
"step": 1600
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 2.992957746478873,
|
| 397 |
+
"grad_norm": 5.347113609313965,
|
| 398 |
+
"learning_rate": 3.892801251956182e-05,
|
| 399 |
+
"loss": 0.2055,
|
| 400 |
+
"step": 1700
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 2.992957746478873,
|
| 404 |
+
"eval_action_accuracy": 0.90625,
|
| 405 |
+
"eval_action_f1": 0.813321647677476,
|
| 406 |
+
"eval_action_precision": 0.8689138576779026,
|
| 407 |
+
"eval_action_recall": 0.7644151565074135,
|
| 408 |
+
"eval_combined_f1": 0.7792135732932114,
|
| 409 |
+
"eval_loss": 1.0456966161727905,
|
| 410 |
+
"eval_runtime": 53.4219,
|
| 411 |
+
"eval_samples_per_second": 42.529,
|
| 412 |
+
"eval_severity_accuracy": 0.7447183098591549,
|
| 413 |
+
"eval_severity_f1": 0.7451054989089468,
|
| 414 |
+
"eval_severity_precision": 0.7459307596916318,
|
| 415 |
+
"eval_severity_recall": 0.7447183098591549,
|
| 416 |
+
"eval_steps_per_second": 2.658,
|
| 417 |
+
"step": 1700
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 3.169014084507042,
|
| 421 |
+
"grad_norm": 8.061402320861816,
|
| 422 |
+
"learning_rate": 3.7949921752738656e-05,
|
| 423 |
+
"loss": 0.1559,
|
| 424 |
+
"step": 1800
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 3.169014084507042,
|
| 428 |
+
"eval_action_accuracy": 0.9049295774647887,
|
| 429 |
+
"eval_action_f1": 0.8036363636363636,
|
| 430 |
+
"eval_action_precision": 0.896551724137931,
|
| 431 |
+
"eval_action_recall": 0.728171334431631,
|
| 432 |
+
"eval_combined_f1": 0.7705730204982484,
|
| 433 |
+
"eval_loss": 1.3266334533691406,
|
| 434 |
+
"eval_runtime": 53.3935,
|
| 435 |
+
"eval_samples_per_second": 42.552,
|
| 436 |
+
"eval_severity_accuracy": 0.7376760563380281,
|
| 437 |
+
"eval_severity_f1": 0.7375096773601331,
|
| 438 |
+
"eval_severity_precision": 0.7543615150494364,
|
| 439 |
+
"eval_severity_recall": 0.7376760563380281,
|
| 440 |
+
"eval_steps_per_second": 2.659,
|
| 441 |
+
"step": 1800
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 3.345070422535211,
|
| 445 |
+
"grad_norm": 4.052455902099609,
|
| 446 |
+
"learning_rate": 3.69718309859155e-05,
|
| 447 |
+
"loss": 0.1174,
|
| 448 |
+
"step": 1900
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 3.345070422535211,
|
| 452 |
+
"eval_action_accuracy": 0.8992077464788732,
|
| 453 |
+
"eval_action_f1": 0.7893284268629255,
|
| 454 |
+
"eval_action_precision": 0.89375,
|
| 455 |
+
"eval_action_recall": 0.7067545304777595,
|
| 456 |
+
"eval_combined_f1": 0.7628428370446565,
|
| 457 |
+
"eval_loss": 1.4744161367416382,
|
| 458 |
+
"eval_runtime": 53.3995,
|
| 459 |
+
"eval_samples_per_second": 42.547,
|
| 460 |
+
"eval_severity_accuracy": 0.7376760563380281,
|
| 461 |
+
"eval_severity_f1": 0.7363572472263874,
|
| 462 |
+
"eval_severity_precision": 0.7497424115719165,
|
| 463 |
+
"eval_severity_recall": 0.7376760563380281,
|
| 464 |
+
"eval_steps_per_second": 2.659,
|
| 465 |
+
"step": 1900
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 3.52112676056338,
|
| 469 |
+
"grad_norm": 6.726060390472412,
|
| 470 |
+
"learning_rate": 3.599374021909233e-05,
|
| 471 |
+
"loss": 0.1325,
|
| 472 |
+
"step": 2000
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 3.52112676056338,
|
| 476 |
+
"eval_action_accuracy": 0.9044894366197183,
|
| 477 |
+
"eval_action_f1": 0.8021877848678214,
|
| 478 |
+
"eval_action_precision": 0.8979591836734694,
|
| 479 |
+
"eval_action_recall": 0.7248764415156508,
|
| 480 |
+
"eval_combined_f1": 0.7739176088843763,
|
| 481 |
+
"eval_loss": 1.592252492904663,
|
| 482 |
+
"eval_runtime": 53.4078,
|
| 483 |
+
"eval_samples_per_second": 42.541,
|
| 484 |
+
"eval_severity_accuracy": 0.7460387323943662,
|
| 485 |
+
"eval_severity_f1": 0.7456474329009313,
|
| 486 |
+
"eval_severity_precision": 0.7468342652042663,
|
| 487 |
+
"eval_severity_recall": 0.7460387323943662,
|
| 488 |
+
"eval_steps_per_second": 2.659,
|
| 489 |
+
"step": 2000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 3.697183098591549,
|
| 493 |
+
"grad_norm": 1.6415315866470337,
|
| 494 |
+
"learning_rate": 3.5015649452269174e-05,
|
| 495 |
+
"loss": 0.1318,
|
| 496 |
+
"step": 2100
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 3.697183098591549,
|
| 500 |
+
"eval_action_accuracy": 0.909330985915493,
|
| 501 |
+
"eval_action_f1": 0.815742397137746,
|
| 502 |
+
"eval_action_precision": 0.8923679060665362,
|
| 503 |
+
"eval_action_recall": 0.7512355848434926,
|
| 504 |
+
"eval_combined_f1": 0.7744664753736676,
|
| 505 |
+
"eval_loss": 1.3085289001464844,
|
| 506 |
+
"eval_runtime": 53.3948,
|
| 507 |
+
"eval_samples_per_second": 42.551,
|
| 508 |
+
"eval_severity_accuracy": 0.7345950704225352,
|
| 509 |
+
"eval_severity_f1": 0.7331905536095894,
|
| 510 |
+
"eval_severity_precision": 0.7463951981865823,
|
| 511 |
+
"eval_severity_recall": 0.7345950704225352,
|
| 512 |
+
"eval_steps_per_second": 2.659,
|
| 513 |
+
"step": 2100
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 3.873239436619718,
|
| 517 |
+
"grad_norm": 1.168869137763977,
|
| 518 |
+
"learning_rate": 3.4037558685446016e-05,
|
| 519 |
+
"loss": 0.1303,
|
| 520 |
+
"step": 2200
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 3.873239436619718,
|
| 524 |
+
"eval_action_accuracy": 0.9080105633802817,
|
| 525 |
+
"eval_action_f1": 0.8271298593879239,
|
| 526 |
+
"eval_action_precision": 0.8305647840531561,
|
| 527 |
+
"eval_action_recall": 0.8237232289950577,
|
| 528 |
+
"eval_combined_f1": 0.7891663886796052,
|
| 529 |
+
"eval_loss": 1.2526799440383911,
|
| 530 |
+
"eval_runtime": 53.3936,
|
| 531 |
+
"eval_samples_per_second": 42.552,
|
| 532 |
+
"eval_severity_accuracy": 0.7513204225352113,
|
| 533 |
+
"eval_severity_f1": 0.7512029179712866,
|
| 534 |
+
"eval_severity_precision": 0.755462312536582,
|
| 535 |
+
"eval_severity_recall": 0.7513204225352113,
|
| 536 |
+
"eval_steps_per_second": 2.659,
|
| 537 |
+
"step": 2200
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.049295774647887,
|
| 541 |
+
"grad_norm": 1.3796656131744385,
|
| 542 |
+
"learning_rate": 3.305946791862285e-05,
|
| 543 |
+
"loss": 0.1226,
|
| 544 |
+
"step": 2300
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 4.049295774647887,
|
| 548 |
+
"eval_action_accuracy": 0.9163732394366197,
|
| 549 |
+
"eval_action_f1": 0.8403361344537815,
|
| 550 |
+
"eval_action_precision": 0.8576329331046312,
|
| 551 |
+
"eval_action_recall": 0.8237232289950577,
|
| 552 |
+
"eval_combined_f1": 0.7939814504731293,
|
| 553 |
+
"eval_loss": 1.3127936124801636,
|
| 554 |
+
"eval_runtime": 53.3969,
|
| 555 |
+
"eval_samples_per_second": 42.549,
|
| 556 |
+
"eval_severity_accuracy": 0.7508802816901409,
|
| 557 |
+
"eval_severity_f1": 0.747626766492477,
|
| 558 |
+
"eval_severity_precision": 0.7590158294773308,
|
| 559 |
+
"eval_severity_recall": 0.7508802816901409,
|
| 560 |
+
"eval_steps_per_second": 2.659,
|
| 561 |
+
"step": 2300
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 4.225352112676056,
|
| 565 |
+
"grad_norm": 4.387660503387451,
|
| 566 |
+
"learning_rate": 3.208137715179969e-05,
|
| 567 |
+
"loss": 0.0703,
|
| 568 |
+
"step": 2400
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 4.225352112676056,
|
| 572 |
+
"eval_action_accuracy": 0.9119718309859155,
|
| 573 |
+
"eval_action_f1": 0.8242530755711776,
|
| 574 |
+
"eval_action_precision": 0.8832391713747646,
|
| 575 |
+
"eval_action_recall": 0.7726523887973641,
|
| 576 |
+
"eval_combined_f1": 0.7795952339344467,
|
| 577 |
+
"eval_loss": 1.73722505569458,
|
| 578 |
+
"eval_runtime": 53.3612,
|
| 579 |
+
"eval_samples_per_second": 42.578,
|
| 580 |
+
"eval_severity_accuracy": 0.7372359154929577,
|
| 581 |
+
"eval_severity_f1": 0.7349373922977159,
|
| 582 |
+
"eval_severity_precision": 0.7478231277482102,
|
| 583 |
+
"eval_severity_recall": 0.7372359154929577,
|
| 584 |
+
"eval_steps_per_second": 2.661,
|
| 585 |
+
"step": 2400
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 4.401408450704225,
|
| 589 |
+
"grad_norm": 3.3632314205169678,
|
| 590 |
+
"learning_rate": 3.110328638497653e-05,
|
| 591 |
+
"loss": 0.0749,
|
| 592 |
+
"step": 2500
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 4.401408450704225,
|
| 596 |
+
"eval_action_accuracy": 0.9058098591549296,
|
| 597 |
+
"eval_action_f1": 0.8145580589254766,
|
| 598 |
+
"eval_action_precision": 0.8592321755027422,
|
| 599 |
+
"eval_action_recall": 0.7742998352553542,
|
| 600 |
+
"eval_combined_f1": 0.7737646532726048,
|
| 601 |
+
"eval_loss": 1.8485854864120483,
|
| 602 |
+
"eval_runtime": 53.3448,
|
| 603 |
+
"eval_samples_per_second": 42.591,
|
| 604 |
+
"eval_severity_accuracy": 0.7319542253521126,
|
| 605 |
+
"eval_severity_f1": 0.7329712476197329,
|
| 606 |
+
"eval_severity_precision": 0.7371728830594799,
|
| 607 |
+
"eval_severity_recall": 0.7319542253521126,
|
| 608 |
+
"eval_steps_per_second": 2.662,
|
| 609 |
+
"step": 2500
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 4.577464788732394,
|
| 613 |
+
"grad_norm": 4.060558319091797,
|
| 614 |
+
"learning_rate": 3.0125195618153366e-05,
|
| 615 |
+
"loss": 0.0776,
|
| 616 |
+
"step": 2600
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 4.577464788732394,
|
| 620 |
+
"eval_action_accuracy": 0.8877640845070423,
|
| 621 |
+
"eval_action_f1": 0.746772591857001,
|
| 622 |
+
"eval_action_precision": 0.94,
|
| 623 |
+
"eval_action_recall": 0.6194398682042833,
|
| 624 |
+
"eval_combined_f1": 0.7451658251590665,
|
| 625 |
+
"eval_loss": 2.1151413917541504,
|
| 626 |
+
"eval_runtime": 53.345,
|
| 627 |
+
"eval_samples_per_second": 42.591,
|
| 628 |
+
"eval_severity_accuracy": 0.7438380281690141,
|
| 629 |
+
"eval_severity_f1": 0.7435590584611321,
|
| 630 |
+
"eval_severity_precision": 0.7627996939449937,
|
| 631 |
+
"eval_severity_recall": 0.7438380281690141,
|
| 632 |
+
"eval_steps_per_second": 2.662,
|
| 633 |
+
"step": 2600
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 4.753521126760563,
|
| 637 |
+
"grad_norm": 1.934782862663269,
|
| 638 |
+
"learning_rate": 2.9147104851330205e-05,
|
| 639 |
+
"loss": 0.0842,
|
| 640 |
+
"step": 2700
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 4.753521126760563,
|
| 644 |
+
"eval_action_accuracy": 0.9154929577464789,
|
| 645 |
+
"eval_action_f1": 0.8449111470113085,
|
| 646 |
+
"eval_action_precision": 0.8288431061806656,
|
| 647 |
+
"eval_action_recall": 0.8616144975288303,
|
| 648 |
+
"eval_combined_f1": 0.7964890047427877,
|
| 649 |
+
"eval_loss": 1.6187307834625244,
|
| 650 |
+
"eval_runtime": 53.3536,
|
| 651 |
+
"eval_samples_per_second": 42.584,
|
| 652 |
+
"eval_severity_accuracy": 0.7477992957746479,
|
| 653 |
+
"eval_severity_f1": 0.748066862474267,
|
| 654 |
+
"eval_severity_precision": 0.751069587929407,
|
| 655 |
+
"eval_severity_recall": 0.7477992957746479,
|
| 656 |
+
"eval_steps_per_second": 2.661,
|
| 657 |
+
"step": 2700
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 4.929577464788732,
|
| 661 |
+
"grad_norm": 0.15578486025333405,
|
| 662 |
+
"learning_rate": 2.8169014084507046e-05,
|
| 663 |
+
"loss": 0.0718,
|
| 664 |
+
"step": 2800
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 4.929577464788732,
|
| 668 |
+
"eval_action_accuracy": 0.9080105633802817,
|
| 669 |
+
"eval_action_f1": 0.8125560538116592,
|
| 670 |
+
"eval_action_precision": 0.8917322834645669,
|
| 671 |
+
"eval_action_recall": 0.7462932454695222,
|
| 672 |
+
"eval_combined_f1": 0.7831897220338088,
|
| 673 |
+
"eval_loss": 1.7883614301681519,
|
| 674 |
+
"eval_runtime": 53.3648,
|
| 675 |
+
"eval_samples_per_second": 42.575,
|
| 676 |
+
"eval_severity_accuracy": 0.7535211267605634,
|
| 677 |
+
"eval_severity_f1": 0.7538233902559583,
|
| 678 |
+
"eval_severity_precision": 0.7606986906314367,
|
| 679 |
+
"eval_severity_recall": 0.7535211267605634,
|
| 680 |
+
"eval_steps_per_second": 2.661,
|
| 681 |
+
"step": 2800
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 5.105633802816901,
|
| 685 |
+
"grad_norm": 0.7947894930839539,
|
| 686 |
+
"learning_rate": 2.719092331768388e-05,
|
| 687 |
+
"loss": 0.0732,
|
| 688 |
+
"step": 2900
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 5.105633802816901,
|
| 692 |
+
"eval_action_accuracy": 0.909330985915493,
|
| 693 |
+
"eval_action_f1": 0.83003300330033,
|
| 694 |
+
"eval_action_precision": 0.8314049586776859,
|
| 695 |
+
"eval_action_recall": 0.828665568369028,
|
| 696 |
+
"eval_combined_f1": 0.7889699480003767,
|
| 697 |
+
"eval_loss": 1.63761305809021,
|
| 698 |
+
"eval_runtime": 53.3715,
|
| 699 |
+
"eval_samples_per_second": 42.57,
|
| 700 |
+
"eval_severity_accuracy": 0.7473591549295775,
|
| 701 |
+
"eval_severity_f1": 0.7479068927004233,
|
| 702 |
+
"eval_severity_precision": 0.7576437021292298,
|
| 703 |
+
"eval_severity_recall": 0.7473591549295775,
|
| 704 |
+
"eval_steps_per_second": 2.661,
|
| 705 |
+
"step": 2900
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 5.28169014084507,
|
| 709 |
+
"grad_norm": 6.216972827911377,
|
| 710 |
+
"learning_rate": 2.6212832550860723e-05,
|
| 711 |
+
"loss": 0.0433,
|
| 712 |
+
"step": 3000
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 5.28169014084507,
|
| 716 |
+
"eval_action_accuracy": 0.9119718309859155,
|
| 717 |
+
"eval_action_f1": 0.823943661971831,
|
| 718 |
+
"eval_action_precision": 0.8846880907372401,
|
| 719 |
+
"eval_action_recall": 0.771004942339374,
|
| 720 |
+
"eval_combined_f1": 0.7916775895426225,
|
| 721 |
+
"eval_loss": 2.1584179401397705,
|
| 722 |
+
"eval_runtime": 53.3702,
|
| 723 |
+
"eval_samples_per_second": 42.571,
|
| 724 |
+
"eval_severity_accuracy": 0.7588028169014085,
|
| 725 |
+
"eval_severity_f1": 0.759411517113414,
|
| 726 |
+
"eval_severity_precision": 0.7667529788427636,
|
| 727 |
+
"eval_severity_recall": 0.7588028169014085,
|
| 728 |
+
"eval_steps_per_second": 2.661,
|
| 729 |
+
"step": 3000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 5.457746478873239,
|
| 733 |
+
"grad_norm": 7.994290351867676,
|
| 734 |
+
"learning_rate": 2.5234741784037558e-05,
|
| 735 |
+
"loss": 0.0437,
|
| 736 |
+
"step": 3100
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 5.457746478873239,
|
| 740 |
+
"eval_action_accuracy": 0.9132922535211268,
|
| 741 |
+
"eval_action_f1": 0.8365145228215768,
|
| 742 |
+
"eval_action_precision": 0.842809364548495,
|
| 743 |
+
"eval_action_recall": 0.8303130148270181,
|
| 744 |
+
"eval_combined_f1": 0.7970826496584587,
|
| 745 |
+
"eval_loss": 1.9302828311920166,
|
| 746 |
+
"eval_runtime": 53.3908,
|
| 747 |
+
"eval_samples_per_second": 42.554,
|
| 748 |
+
"eval_severity_accuracy": 0.758362676056338,
|
| 749 |
+
"eval_severity_f1": 0.7576507764953404,
|
| 750 |
+
"eval_severity_precision": 0.7661980956912092,
|
| 751 |
+
"eval_severity_recall": 0.758362676056338,
|
| 752 |
+
"eval_steps_per_second": 2.66,
|
| 753 |
+
"step": 3100
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 5.633802816901408,
|
| 757 |
+
"grad_norm": 0.0936056599020958,
|
| 758 |
+
"learning_rate": 2.42566510172144e-05,
|
| 759 |
+
"loss": 0.0329,
|
| 760 |
+
"step": 3200
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 5.633802816901408,
|
| 764 |
+
"eval_action_accuracy": 0.9137323943661971,
|
| 765 |
+
"eval_action_f1": 0.833616298811545,
|
| 766 |
+
"eval_action_precision": 0.8598949211908932,
|
| 767 |
+
"eval_action_recall": 0.8088962108731467,
|
| 768 |
+
"eval_combined_f1": 0.7923034119567994,
|
| 769 |
+
"eval_loss": 2.4584481716156006,
|
| 770 |
+
"eval_runtime": 53.4118,
|
| 771 |
+
"eval_samples_per_second": 42.537,
|
| 772 |
+
"eval_severity_accuracy": 0.7522007042253521,
|
| 773 |
+
"eval_severity_f1": 0.7509905251020539,
|
| 774 |
+
"eval_severity_precision": 0.7646859410918823,
|
| 775 |
+
"eval_severity_recall": 0.7522007042253521,
|
| 776 |
+
"eval_steps_per_second": 2.659,
|
| 777 |
+
"step": 3200
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 5.809859154929578,
|
| 781 |
+
"grad_norm": 0.029598630964756012,
|
| 782 |
+
"learning_rate": 2.327856025039124e-05,
|
| 783 |
+
"loss": 0.0566,
|
| 784 |
+
"step": 3300
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 5.809859154929578,
|
| 788 |
+
"eval_action_accuracy": 0.9132922535211268,
|
| 789 |
+
"eval_action_f1": 0.841000807102502,
|
| 790 |
+
"eval_action_precision": 0.8243670886075949,
|
| 791 |
+
"eval_action_recall": 0.8583196046128501,
|
| 792 |
+
"eval_combined_f1": 0.8000965798413886,
|
| 793 |
+
"eval_loss": 1.6451172828674316,
|
| 794 |
+
"eval_runtime": 53.3855,
|
| 795 |
+
"eval_samples_per_second": 42.558,
|
| 796 |
+
"eval_severity_accuracy": 0.7588028169014085,
|
| 797 |
+
"eval_severity_f1": 0.7591923525802752,
|
| 798 |
+
"eval_severity_precision": 0.7658252855908393,
|
| 799 |
+
"eval_severity_recall": 0.7588028169014085,
|
| 800 |
+
"eval_steps_per_second": 2.66,
|
| 801 |
+
"step": 3300
|
| 802 |
+
}
|
| 803 |
+
],
|
| 804 |
+
"logging_steps": 100,
|
| 805 |
+
"max_steps": 5680,
|
| 806 |
+
"num_input_tokens_seen": 0,
|
| 807 |
+
"num_train_epochs": 10,
|
| 808 |
+
"save_steps": 100,
|
| 809 |
+
"stateful_callbacks": {
|
| 810 |
+
"TrainerControl": {
|
| 811 |
+
"args": {
|
| 812 |
+
"should_epoch_stop": false,
|
| 813 |
+
"should_evaluate": false,
|
| 814 |
+
"should_log": false,
|
| 815 |
+
"should_save": true,
|
| 816 |
+
"should_training_stop": false
|
| 817 |
+
},
|
| 818 |
+
"attributes": {}
|
| 819 |
+
}
|
| 820 |
+
},
|
| 821 |
+
"total_flos": 5.581830484199424e+16,
|
| 822 |
+
"train_batch_size": 16,
|
| 823 |
+
"trial_name": null,
|
| 824 |
+
"trial_params": null
|
| 825 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c551536363f8fa7fc4d291ae6e7bfe434eea3de0cd497fac23469563e4084824
|
| 3 |
+
size 5368
|