Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/trainer_state.json +345 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/trainer_state.json +441 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/trainer_state.json +235 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/trainer_state.json +331 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/trainer_state.json +263 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/trainer_state.json +380 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/trainer_state.json +217 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/trainer_state.json +327 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/trainer_state.json +473 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/trainer_state.json +377 -0
- ood-viclickbait2025-bamibert-source-source-05-s43-20260827T173648Z-32d744c6/trainer/checkpoint-810/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-05-s43-20260827T173648Z-32d744c6/trainer/checkpoint-810/trainer_state.json +366 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/trainer_state.json +224 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/special_tokens_map.json +51 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/tokenizer.json +0 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/tokenizer_config.json +55 -0
- ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/trainer_state.json +327 -0
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729/trainer_state.json
ADDED
|
@@ -0,0 +1,345 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 729,
|
| 3 |
+
"best_metric": 0.8309502997002998,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 729,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.3115264797507788,
|
| 14 |
+
"grad_norm": 20.875186920166016,
|
| 15 |
+
"learning_rate": 2.962962962962963e-06,
|
| 16 |
+
"loss": 2.5336,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6230529595015576,
|
| 21 |
+
"grad_norm": 10.69560432434082,
|
| 22 |
+
"learning_rate": 6.049382716049383e-06,
|
| 23 |
+
"loss": 2.4469,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9345794392523364,
|
| 28 |
+
"grad_norm": 94.32146453857422,
|
| 29 |
+
"learning_rate": 9.135802469135803e-06,
|
| 30 |
+
"loss": 2.3862,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.7368421052631579,
|
| 36 |
+
"eval_loss": 0.5407072901725769,
|
| 37 |
+
"eval_macro_f1": 0.42424242424242425,
|
| 38 |
+
"eval_runtime": 0.2967,
|
| 39 |
+
"eval_samples_per_second": 960.548,
|
| 40 |
+
"eval_steps_per_second": 60.666,
|
| 41 |
+
"eval_weighted_f1": 0.6251993620414673,
|
| 42 |
+
"step": 81
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.236760124610592,
|
| 46 |
+
"grad_norm": 16.57114028930664,
|
| 47 |
+
"learning_rate": 1.2222222222222224e-05,
|
| 48 |
+
"loss": 2.1731,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.5482866043613708,
|
| 53 |
+
"grad_norm": 25.705949783325195,
|
| 54 |
+
"learning_rate": 1.5308641975308643e-05,
|
| 55 |
+
"loss": 1.8767,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.8598130841121496,
|
| 60 |
+
"grad_norm": 32.91946792602539,
|
| 61 |
+
"learning_rate": 1.8395061728395062e-05,
|
| 62 |
+
"loss": 1.7056,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.7929824561403509,
|
| 68 |
+
"eval_loss": 0.4607318937778473,
|
| 69 |
+
"eval_macro_f1": 0.6827418350597159,
|
| 70 |
+
"eval_runtime": 0.272,
|
| 71 |
+
"eval_samples_per_second": 1047.795,
|
| 72 |
+
"eval_steps_per_second": 66.177,
|
| 73 |
+
"eval_weighted_f1": 0.7713280484280833,
|
| 74 |
+
"step": 162
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.161993769470405,
|
| 78 |
+
"grad_norm": 16.986635208129883,
|
| 79 |
+
"learning_rate": 1.983539094650206e-05,
|
| 80 |
+
"loss": 1.5277,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.473520249221184,
|
| 85 |
+
"grad_norm": 21.75818634033203,
|
| 86 |
+
"learning_rate": 1.9492455418381347e-05,
|
| 87 |
+
"loss": 1.4474,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.7850467289719627,
|
| 92 |
+
"grad_norm": 48.18912124633789,
|
| 93 |
+
"learning_rate": 1.9149519890260632e-05,
|
| 94 |
+
"loss": 1.2291,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8526315789473684,
|
| 100 |
+
"eval_loss": 0.3137834668159485,
|
| 101 |
+
"eval_macro_f1": 0.814659977703456,
|
| 102 |
+
"eval_runtime": 0.2536,
|
| 103 |
+
"eval_samples_per_second": 1123.87,
|
| 104 |
+
"eval_steps_per_second": 70.981,
|
| 105 |
+
"eval_weighted_f1": 0.8543976999354573,
|
| 106 |
+
"step": 243
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.087227414330218,
|
| 110 |
+
"grad_norm": 37.558712005615234,
|
| 111 |
+
"learning_rate": 1.8806584362139917e-05,
|
| 112 |
+
"loss": 1.157,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.398753894080997,
|
| 117 |
+
"grad_norm": 92.24092102050781,
|
| 118 |
+
"learning_rate": 1.8463648834019206e-05,
|
| 119 |
+
"loss": 1.0507,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.710280373831776,
|
| 124 |
+
"grad_norm": 37.75823211669922,
|
| 125 |
+
"learning_rate": 1.812071330589849e-05,
|
| 126 |
+
"loss": 1.0149,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8526315789473684,
|
| 132 |
+
"eval_loss": 0.33415332436561584,
|
| 133 |
+
"eval_macro_f1": 0.8188997821350763,
|
| 134 |
+
"eval_runtime": 0.2608,
|
| 135 |
+
"eval_samples_per_second": 1092.828,
|
| 136 |
+
"eval_steps_per_second": 69.021,
|
| 137 |
+
"eval_weighted_f1": 0.8559224859534457,
|
| 138 |
+
"step": 324
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.012461059190031,
|
| 142 |
+
"grad_norm": 17.798608779907227,
|
| 143 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 144 |
+
"loss": 1.1308,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.32398753894081,
|
| 149 |
+
"grad_norm": 56.464027404785156,
|
| 150 |
+
"learning_rate": 1.7434842249657065e-05,
|
| 151 |
+
"loss": 0.7207,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.635514018691588,
|
| 156 |
+
"grad_norm": 66.06902313232422,
|
| 157 |
+
"learning_rate": 1.7091906721536354e-05,
|
| 158 |
+
"loss": 0.9485,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.947040498442368,
|
| 163 |
+
"grad_norm": 40.791996002197266,
|
| 164 |
+
"learning_rate": 1.674897119341564e-05,
|
| 165 |
+
"loss": 0.8455,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 5.0,
|
| 170 |
+
"eval_accuracy": 0.8631578947368421,
|
| 171 |
+
"eval_loss": 0.39822354912757874,
|
| 172 |
+
"eval_macro_f1": 0.8087610330173258,
|
| 173 |
+
"eval_runtime": 0.2772,
|
| 174 |
+
"eval_samples_per_second": 1027.964,
|
| 175 |
+
"eval_steps_per_second": 64.924,
|
| 176 |
+
"eval_weighted_f1": 0.8570740352024225,
|
| 177 |
+
"step": 405
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.249221183800623,
|
| 181 |
+
"grad_norm": 73.6117935180664,
|
| 182 |
+
"learning_rate": 1.6406035665294924e-05,
|
| 183 |
+
"loss": 0.5354,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.5607476635514015,
|
| 188 |
+
"grad_norm": 76.78546905517578,
|
| 189 |
+
"learning_rate": 1.6063100137174213e-05,
|
| 190 |
+
"loss": 0.6269,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.872274143302181,
|
| 195 |
+
"grad_norm": 1.4672813415527344,
|
| 196 |
+
"learning_rate": 1.5720164609053498e-05,
|
| 197 |
+
"loss": 0.4834,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 6.0,
|
| 202 |
+
"eval_accuracy": 0.8596491228070176,
|
| 203 |
+
"eval_loss": 0.49580055475234985,
|
| 204 |
+
"eval_macro_f1": 0.819047619047619,
|
| 205 |
+
"eval_runtime": 0.2651,
|
| 206 |
+
"eval_samples_per_second": 1074.996,
|
| 207 |
+
"eval_steps_per_second": 67.895,
|
| 208 |
+
"eval_weighted_f1": 0.8596491228070176,
|
| 209 |
+
"step": 486
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.174454828660436,
|
| 213 |
+
"grad_norm": 109.67298126220703,
|
| 214 |
+
"learning_rate": 1.5377229080932784e-05,
|
| 215 |
+
"loss": 0.3795,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 6.485981308411215,
|
| 220 |
+
"grad_norm": 69.39620971679688,
|
| 221 |
+
"learning_rate": 1.5034293552812072e-05,
|
| 222 |
+
"loss": 0.6088,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.797507788161994,
|
| 227 |
+
"grad_norm": 76.5630874633789,
|
| 228 |
+
"learning_rate": 1.469135802469136e-05,
|
| 229 |
+
"loss": 0.4297,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 7.0,
|
| 234 |
+
"eval_accuracy": 0.8350877192982457,
|
| 235 |
+
"eval_loss": 0.7009875774383545,
|
| 236 |
+
"eval_macro_f1": 0.7845804988662131,
|
| 237 |
+
"eval_runtime": 0.2581,
|
| 238 |
+
"eval_samples_per_second": 1104.023,
|
| 239 |
+
"eval_steps_per_second": 69.728,
|
| 240 |
+
"eval_weighted_f1": 0.8339897362453753,
|
| 241 |
+
"step": 567
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.099688473520249,
|
| 245 |
+
"grad_norm": 27.954246520996094,
|
| 246 |
+
"learning_rate": 1.4348422496570646e-05,
|
| 247 |
+
"loss": 0.3056,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 7.411214953271028,
|
| 252 |
+
"grad_norm": 76.87083435058594,
|
| 253 |
+
"learning_rate": 1.4005486968449933e-05,
|
| 254 |
+
"loss": 0.3368,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 7.722741433021807,
|
| 259 |
+
"grad_norm": 0.35162127017974854,
|
| 260 |
+
"learning_rate": 1.366255144032922e-05,
|
| 261 |
+
"loss": 0.3579,
|
| 262 |
+
"step": 625
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 8.0,
|
| 266 |
+
"eval_accuracy": 0.8280701754385965,
|
| 267 |
+
"eval_loss": 0.9167160391807556,
|
| 268 |
+
"eval_macro_f1": 0.7460031647296339,
|
| 269 |
+
"eval_runtime": 0.2592,
|
| 270 |
+
"eval_samples_per_second": 1099.542,
|
| 271 |
+
"eval_steps_per_second": 69.445,
|
| 272 |
+
"eval_weighted_f1": 0.8143923403204361,
|
| 273 |
+
"step": 648
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.024922118380061,
|
| 277 |
+
"grad_norm": 45.39617919921875,
|
| 278 |
+
"learning_rate": 1.3319615912208505e-05,
|
| 279 |
+
"loss": 0.1946,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 8.336448598130842,
|
| 284 |
+
"grad_norm": 16.47720718383789,
|
| 285 |
+
"learning_rate": 1.2976680384087792e-05,
|
| 286 |
+
"loss": 0.3102,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 8.64797507788162,
|
| 291 |
+
"grad_norm": 0.5873615145683289,
|
| 292 |
+
"learning_rate": 1.2633744855967079e-05,
|
| 293 |
+
"loss": 0.2386,
|
| 294 |
+
"step": 700
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 8.959501557632398,
|
| 298 |
+
"grad_norm": 3.4351816177368164,
|
| 299 |
+
"learning_rate": 1.2290809327846366e-05,
|
| 300 |
+
"loss": 0.2555,
|
| 301 |
+
"step": 725
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 9.0,
|
| 305 |
+
"eval_accuracy": 0.8666666666666667,
|
| 306 |
+
"eval_loss": 0.8016504049301147,
|
| 307 |
+
"eval_macro_f1": 0.8309502997002998,
|
| 308 |
+
"eval_runtime": 0.2806,
|
| 309 |
+
"eval_samples_per_second": 1015.655,
|
| 310 |
+
"eval_steps_per_second": 64.147,
|
| 311 |
+
"eval_weighted_f1": 0.8677572427572426,
|
| 312 |
+
"step": 729
|
| 313 |
+
}
|
| 314 |
+
],
|
| 315 |
+
"logging_steps": 25,
|
| 316 |
+
"max_steps": 1620,
|
| 317 |
+
"num_input_tokens_seen": 0,
|
| 318 |
+
"num_train_epochs": 20,
|
| 319 |
+
"save_steps": 500,
|
| 320 |
+
"stateful_callbacks": {
|
| 321 |
+
"EarlyStoppingCallback": {
|
| 322 |
+
"args": {
|
| 323 |
+
"early_stopping_patience": 3,
|
| 324 |
+
"early_stopping_threshold": 0.0
|
| 325 |
+
},
|
| 326 |
+
"attributes": {
|
| 327 |
+
"early_stopping_patience_counter": 0
|
| 328 |
+
}
|
| 329 |
+
},
|
| 330 |
+
"TrainerControl": {
|
| 331 |
+
"args": {
|
| 332 |
+
"should_epoch_stop": false,
|
| 333 |
+
"should_evaluate": false,
|
| 334 |
+
"should_log": false,
|
| 335 |
+
"should_save": true,
|
| 336 |
+
"should_training_stop": false
|
| 337 |
+
},
|
| 338 |
+
"attributes": {}
|
| 339 |
+
}
|
| 340 |
+
},
|
| 341 |
+
"total_flos": 0.0,
|
| 342 |
+
"train_batch_size": 8,
|
| 343 |
+
"trial_name": null,
|
| 344 |
+
"trial_params": null
|
| 345 |
+
}
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-972/trainer_state.json
ADDED
|
@@ -0,0 +1,441 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 729,
|
| 3 |
+
"best_metric": 0.8309502997002998,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-00-s44-20260827T174134Z-fe731097/trainer/checkpoint-729",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 972,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.3115264797507788,
|
| 14 |
+
"grad_norm": 20.875186920166016,
|
| 15 |
+
"learning_rate": 2.962962962962963e-06,
|
| 16 |
+
"loss": 2.5336,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6230529595015576,
|
| 21 |
+
"grad_norm": 10.69560432434082,
|
| 22 |
+
"learning_rate": 6.049382716049383e-06,
|
| 23 |
+
"loss": 2.4469,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9345794392523364,
|
| 28 |
+
"grad_norm": 94.32146453857422,
|
| 29 |
+
"learning_rate": 9.135802469135803e-06,
|
| 30 |
+
"loss": 2.3862,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.7368421052631579,
|
| 36 |
+
"eval_loss": 0.5407072901725769,
|
| 37 |
+
"eval_macro_f1": 0.42424242424242425,
|
| 38 |
+
"eval_runtime": 0.2967,
|
| 39 |
+
"eval_samples_per_second": 960.548,
|
| 40 |
+
"eval_steps_per_second": 60.666,
|
| 41 |
+
"eval_weighted_f1": 0.6251993620414673,
|
| 42 |
+
"step": 81
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.236760124610592,
|
| 46 |
+
"grad_norm": 16.57114028930664,
|
| 47 |
+
"learning_rate": 1.2222222222222224e-05,
|
| 48 |
+
"loss": 2.1731,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.5482866043613708,
|
| 53 |
+
"grad_norm": 25.705949783325195,
|
| 54 |
+
"learning_rate": 1.5308641975308643e-05,
|
| 55 |
+
"loss": 1.8767,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.8598130841121496,
|
| 60 |
+
"grad_norm": 32.91946792602539,
|
| 61 |
+
"learning_rate": 1.8395061728395062e-05,
|
| 62 |
+
"loss": 1.7056,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.7929824561403509,
|
| 68 |
+
"eval_loss": 0.4607318937778473,
|
| 69 |
+
"eval_macro_f1": 0.6827418350597159,
|
| 70 |
+
"eval_runtime": 0.272,
|
| 71 |
+
"eval_samples_per_second": 1047.795,
|
| 72 |
+
"eval_steps_per_second": 66.177,
|
| 73 |
+
"eval_weighted_f1": 0.7713280484280833,
|
| 74 |
+
"step": 162
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.161993769470405,
|
| 78 |
+
"grad_norm": 16.986635208129883,
|
| 79 |
+
"learning_rate": 1.983539094650206e-05,
|
| 80 |
+
"loss": 1.5277,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.473520249221184,
|
| 85 |
+
"grad_norm": 21.75818634033203,
|
| 86 |
+
"learning_rate": 1.9492455418381347e-05,
|
| 87 |
+
"loss": 1.4474,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.7850467289719627,
|
| 92 |
+
"grad_norm": 48.18912124633789,
|
| 93 |
+
"learning_rate": 1.9149519890260632e-05,
|
| 94 |
+
"loss": 1.2291,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8526315789473684,
|
| 100 |
+
"eval_loss": 0.3137834668159485,
|
| 101 |
+
"eval_macro_f1": 0.814659977703456,
|
| 102 |
+
"eval_runtime": 0.2536,
|
| 103 |
+
"eval_samples_per_second": 1123.87,
|
| 104 |
+
"eval_steps_per_second": 70.981,
|
| 105 |
+
"eval_weighted_f1": 0.8543976999354573,
|
| 106 |
+
"step": 243
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.087227414330218,
|
| 110 |
+
"grad_norm": 37.558712005615234,
|
| 111 |
+
"learning_rate": 1.8806584362139917e-05,
|
| 112 |
+
"loss": 1.157,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.398753894080997,
|
| 117 |
+
"grad_norm": 92.24092102050781,
|
| 118 |
+
"learning_rate": 1.8463648834019206e-05,
|
| 119 |
+
"loss": 1.0507,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.710280373831776,
|
| 124 |
+
"grad_norm": 37.75823211669922,
|
| 125 |
+
"learning_rate": 1.812071330589849e-05,
|
| 126 |
+
"loss": 1.0149,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8526315789473684,
|
| 132 |
+
"eval_loss": 0.33415332436561584,
|
| 133 |
+
"eval_macro_f1": 0.8188997821350763,
|
| 134 |
+
"eval_runtime": 0.2608,
|
| 135 |
+
"eval_samples_per_second": 1092.828,
|
| 136 |
+
"eval_steps_per_second": 69.021,
|
| 137 |
+
"eval_weighted_f1": 0.8559224859534457,
|
| 138 |
+
"step": 324
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.012461059190031,
|
| 142 |
+
"grad_norm": 17.798608779907227,
|
| 143 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 144 |
+
"loss": 1.1308,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.32398753894081,
|
| 149 |
+
"grad_norm": 56.464027404785156,
|
| 150 |
+
"learning_rate": 1.7434842249657065e-05,
|
| 151 |
+
"loss": 0.7207,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.635514018691588,
|
| 156 |
+
"grad_norm": 66.06902313232422,
|
| 157 |
+
"learning_rate": 1.7091906721536354e-05,
|
| 158 |
+
"loss": 0.9485,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.947040498442368,
|
| 163 |
+
"grad_norm": 40.791996002197266,
|
| 164 |
+
"learning_rate": 1.674897119341564e-05,
|
| 165 |
+
"loss": 0.8455,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 5.0,
|
| 170 |
+
"eval_accuracy": 0.8631578947368421,
|
| 171 |
+
"eval_loss": 0.39822354912757874,
|
| 172 |
+
"eval_macro_f1": 0.8087610330173258,
|
| 173 |
+
"eval_runtime": 0.2772,
|
| 174 |
+
"eval_samples_per_second": 1027.964,
|
| 175 |
+
"eval_steps_per_second": 64.924,
|
| 176 |
+
"eval_weighted_f1": 0.8570740352024225,
|
| 177 |
+
"step": 405
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.249221183800623,
|
| 181 |
+
"grad_norm": 73.6117935180664,
|
| 182 |
+
"learning_rate": 1.6406035665294924e-05,
|
| 183 |
+
"loss": 0.5354,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.5607476635514015,
|
| 188 |
+
"grad_norm": 76.78546905517578,
|
| 189 |
+
"learning_rate": 1.6063100137174213e-05,
|
| 190 |
+
"loss": 0.6269,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.872274143302181,
|
| 195 |
+
"grad_norm": 1.4672813415527344,
|
| 196 |
+
"learning_rate": 1.5720164609053498e-05,
|
| 197 |
+
"loss": 0.4834,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 6.0,
|
| 202 |
+
"eval_accuracy": 0.8596491228070176,
|
| 203 |
+
"eval_loss": 0.49580055475234985,
|
| 204 |
+
"eval_macro_f1": 0.819047619047619,
|
| 205 |
+
"eval_runtime": 0.2651,
|
| 206 |
+
"eval_samples_per_second": 1074.996,
|
| 207 |
+
"eval_steps_per_second": 67.895,
|
| 208 |
+
"eval_weighted_f1": 0.8596491228070176,
|
| 209 |
+
"step": 486
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.174454828660436,
|
| 213 |
+
"grad_norm": 109.67298126220703,
|
| 214 |
+
"learning_rate": 1.5377229080932784e-05,
|
| 215 |
+
"loss": 0.3795,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 6.485981308411215,
|
| 220 |
+
"grad_norm": 69.39620971679688,
|
| 221 |
+
"learning_rate": 1.5034293552812072e-05,
|
| 222 |
+
"loss": 0.6088,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.797507788161994,
|
| 227 |
+
"grad_norm": 76.5630874633789,
|
| 228 |
+
"learning_rate": 1.469135802469136e-05,
|
| 229 |
+
"loss": 0.4297,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 7.0,
|
| 234 |
+
"eval_accuracy": 0.8350877192982457,
|
| 235 |
+
"eval_loss": 0.7009875774383545,
|
| 236 |
+
"eval_macro_f1": 0.7845804988662131,
|
| 237 |
+
"eval_runtime": 0.2581,
|
| 238 |
+
"eval_samples_per_second": 1104.023,
|
| 239 |
+
"eval_steps_per_second": 69.728,
|
| 240 |
+
"eval_weighted_f1": 0.8339897362453753,
|
| 241 |
+
"step": 567
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.099688473520249,
|
| 245 |
+
"grad_norm": 27.954246520996094,
|
| 246 |
+
"learning_rate": 1.4348422496570646e-05,
|
| 247 |
+
"loss": 0.3056,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 7.411214953271028,
|
| 252 |
+
"grad_norm": 76.87083435058594,
|
| 253 |
+
"learning_rate": 1.4005486968449933e-05,
|
| 254 |
+
"loss": 0.3368,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 7.722741433021807,
|
| 259 |
+
"grad_norm": 0.35162127017974854,
|
| 260 |
+
"learning_rate": 1.366255144032922e-05,
|
| 261 |
+
"loss": 0.3579,
|
| 262 |
+
"step": 625
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 8.0,
|
| 266 |
+
"eval_accuracy": 0.8280701754385965,
|
| 267 |
+
"eval_loss": 0.9167160391807556,
|
| 268 |
+
"eval_macro_f1": 0.7460031647296339,
|
| 269 |
+
"eval_runtime": 0.2592,
|
| 270 |
+
"eval_samples_per_second": 1099.542,
|
| 271 |
+
"eval_steps_per_second": 69.445,
|
| 272 |
+
"eval_weighted_f1": 0.8143923403204361,
|
| 273 |
+
"step": 648
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.024922118380061,
|
| 277 |
+
"grad_norm": 45.39617919921875,
|
| 278 |
+
"learning_rate": 1.3319615912208505e-05,
|
| 279 |
+
"loss": 0.1946,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 8.336448598130842,
|
| 284 |
+
"grad_norm": 16.47720718383789,
|
| 285 |
+
"learning_rate": 1.2976680384087792e-05,
|
| 286 |
+
"loss": 0.3102,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 8.64797507788162,
|
| 291 |
+
"grad_norm": 0.5873615145683289,
|
| 292 |
+
"learning_rate": 1.2633744855967079e-05,
|
| 293 |
+
"loss": 0.2386,
|
| 294 |
+
"step": 700
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 8.959501557632398,
|
| 298 |
+
"grad_norm": 3.4351816177368164,
|
| 299 |
+
"learning_rate": 1.2290809327846366e-05,
|
| 300 |
+
"loss": 0.2555,
|
| 301 |
+
"step": 725
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 9.0,
|
| 305 |
+
"eval_accuracy": 0.8666666666666667,
|
| 306 |
+
"eval_loss": 0.8016504049301147,
|
| 307 |
+
"eval_macro_f1": 0.8309502997002998,
|
| 308 |
+
"eval_runtime": 0.2806,
|
| 309 |
+
"eval_samples_per_second": 1015.655,
|
| 310 |
+
"eval_steps_per_second": 64.147,
|
| 311 |
+
"eval_weighted_f1": 0.8677572427572426,
|
| 312 |
+
"step": 729
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 9.261682242990654,
|
| 316 |
+
"grad_norm": 0.7731945514678955,
|
| 317 |
+
"learning_rate": 1.1947873799725653e-05,
|
| 318 |
+
"loss": 0.2872,
|
| 319 |
+
"step": 750
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 9.573208722741432,
|
| 323 |
+
"grad_norm": 82.8419418334961,
|
| 324 |
+
"learning_rate": 1.160493827160494e-05,
|
| 325 |
+
"loss": 0.3674,
|
| 326 |
+
"step": 775
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 9.884735202492212,
|
| 330 |
+
"grad_norm": 57.24100875854492,
|
| 331 |
+
"learning_rate": 1.1262002743484225e-05,
|
| 332 |
+
"loss": 0.1757,
|
| 333 |
+
"step": 800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 10.0,
|
| 337 |
+
"eval_accuracy": 0.8666666666666667,
|
| 338 |
+
"eval_loss": 0.854817271232605,
|
| 339 |
+
"eval_macro_f1": 0.8265979249391571,
|
| 340 |
+
"eval_runtime": 0.2424,
|
| 341 |
+
"eval_samples_per_second": 1175.771,
|
| 342 |
+
"eval_steps_per_second": 74.259,
|
| 343 |
+
"eval_weighted_f1": 0.8660817215319585,
|
| 344 |
+
"step": 810
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 10.186915887850468,
|
| 348 |
+
"grad_norm": 0.030899252742528915,
|
| 349 |
+
"learning_rate": 1.0919067215363512e-05,
|
| 350 |
+
"loss": 0.1444,
|
| 351 |
+
"step": 825
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 10.498442367601246,
|
| 355 |
+
"grad_norm": 0.6250601410865784,
|
| 356 |
+
"learning_rate": 1.0576131687242799e-05,
|
| 357 |
+
"loss": 0.1753,
|
| 358 |
+
"step": 850
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 10.809968847352025,
|
| 362 |
+
"grad_norm": 0.49999409914016724,
|
| 363 |
+
"learning_rate": 1.0233196159122084e-05,
|
| 364 |
+
"loss": 0.2879,
|
| 365 |
+
"step": 875
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 11.0,
|
| 369 |
+
"eval_accuracy": 0.8596491228070176,
|
| 370 |
+
"eval_loss": 0.9294641613960266,
|
| 371 |
+
"eval_macro_f1": 0.7962247962247962,
|
| 372 |
+
"eval_runtime": 0.2647,
|
| 373 |
+
"eval_samples_per_second": 1076.749,
|
| 374 |
+
"eval_steps_per_second": 68.005,
|
| 375 |
+
"eval_weighted_f1": 0.8500756395493239,
|
| 376 |
+
"step": 891
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 11.11214953271028,
|
| 380 |
+
"grad_norm": 0.20545834302902222,
|
| 381 |
+
"learning_rate": 9.890260631001373e-06,
|
| 382 |
+
"loss": 0.2422,
|
| 383 |
+
"step": 900
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 11.423676012461058,
|
| 387 |
+
"grad_norm": 103.7739028930664,
|
| 388 |
+
"learning_rate": 9.547325102880658e-06,
|
| 389 |
+
"loss": 0.0932,
|
| 390 |
+
"step": 925
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 11.735202492211839,
|
| 394 |
+
"grad_norm": 0.26511281728744507,
|
| 395 |
+
"learning_rate": 9.204389574759945e-06,
|
| 396 |
+
"loss": 0.2531,
|
| 397 |
+
"step": 950
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 12.0,
|
| 401 |
+
"eval_accuracy": 0.8526315789473684,
|
| 402 |
+
"eval_loss": 0.9247146248817444,
|
| 403 |
+
"eval_macro_f1": 0.8116028708133971,
|
| 404 |
+
"eval_runtime": 0.2695,
|
| 405 |
+
"eval_samples_per_second": 1057.528,
|
| 406 |
+
"eval_steps_per_second": 66.791,
|
| 407 |
+
"eval_weighted_f1": 0.8532485520020145,
|
| 408 |
+
"step": 972
|
| 409 |
+
}
|
| 410 |
+
],
|
| 411 |
+
"logging_steps": 25,
|
| 412 |
+
"max_steps": 1620,
|
| 413 |
+
"num_input_tokens_seen": 0,
|
| 414 |
+
"num_train_epochs": 20,
|
| 415 |
+
"save_steps": 500,
|
| 416 |
+
"stateful_callbacks": {
|
| 417 |
+
"EarlyStoppingCallback": {
|
| 418 |
+
"args": {
|
| 419 |
+
"early_stopping_patience": 3,
|
| 420 |
+
"early_stopping_threshold": 0.0
|
| 421 |
+
},
|
| 422 |
+
"attributes": {
|
| 423 |
+
"early_stopping_patience_counter": 3
|
| 424 |
+
}
|
| 425 |
+
},
|
| 426 |
+
"TrainerControl": {
|
| 427 |
+
"args": {
|
| 428 |
+
"should_epoch_stop": false,
|
| 429 |
+
"should_evaluate": false,
|
| 430 |
+
"should_log": false,
|
| 431 |
+
"should_save": true,
|
| 432 |
+
"should_training_stop": true
|
| 433 |
+
},
|
| 434 |
+
"attributes": {}
|
| 435 |
+
}
|
| 436 |
+
},
|
| 437 |
+
"total_flos": 0.0,
|
| 438 |
+
"train_batch_size": 8,
|
| 439 |
+
"trial_name": null,
|
| 440 |
+
"trial_params": null
|
| 441 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456/trainer_state.json
ADDED
|
@@ -0,0 +1,235 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 456,
|
| 3 |
+
"best_metric": 0.8209913799774053,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 456,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33112582781456956,
|
| 14 |
+
"grad_norm": 29.215723037719727,
|
| 15 |
+
"learning_rate": 3.157894736842105e-06,
|
| 16 |
+
"loss": 2.6441,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6622516556291391,
|
| 21 |
+
"grad_norm": 10.825380325317383,
|
| 22 |
+
"learning_rate": 6.447368421052632e-06,
|
| 23 |
+
"loss": 2.5416,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9933774834437086,
|
| 28 |
+
"grad_norm": 12.073180198669434,
|
| 29 |
+
"learning_rate": 9.736842105263159e-06,
|
| 30 |
+
"loss": 2.5375,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6468401486988847,
|
| 36 |
+
"eval_loss": 0.6156554222106934,
|
| 37 |
+
"eval_macro_f1": 0.3927765237020316,
|
| 38 |
+
"eval_runtime": 0.2515,
|
| 39 |
+
"eval_samples_per_second": 1069.733,
|
| 40 |
+
"eval_steps_per_second": 67.604,
|
| 41 |
+
"eval_weighted_f1": 0.5081272499937063,
|
| 42 |
+
"step": 76
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.3178807947019868,
|
| 46 |
+
"grad_norm": 31.75727081298828,
|
| 47 |
+
"learning_rate": 1.3026315789473684e-05,
|
| 48 |
+
"loss": 2.2574,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6490066225165563,
|
| 53 |
+
"grad_norm": 29.24384307861328,
|
| 54 |
+
"learning_rate": 1.6315789473684213e-05,
|
| 55 |
+
"loss": 2.0182,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.980132450331126,
|
| 60 |
+
"grad_norm": 45.39990997314453,
|
| 61 |
+
"learning_rate": 1.960526315789474e-05,
|
| 62 |
+
"loss": 1.9235,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.7992565055762082,
|
| 68 |
+
"eval_loss": 0.39099982380867004,
|
| 69 |
+
"eval_macro_f1": 0.773314606741573,
|
| 70 |
+
"eval_runtime": 0.2287,
|
| 71 |
+
"eval_samples_per_second": 1176.414,
|
| 72 |
+
"eval_steps_per_second": 74.346,
|
| 73 |
+
"eval_weighted_f1": 0.7958355958397728,
|
| 74 |
+
"step": 152
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.304635761589404,
|
| 78 |
+
"grad_norm": 58.309898376464844,
|
| 79 |
+
"learning_rate": 1.9678362573099416e-05,
|
| 80 |
+
"loss": 1.5733,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.6357615894039736,
|
| 85 |
+
"grad_norm": 35.05423355102539,
|
| 86 |
+
"learning_rate": 1.9312865497076027e-05,
|
| 87 |
+
"loss": 1.5295,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.966887417218543,
|
| 92 |
+
"grad_norm": 15.614839553833008,
|
| 93 |
+
"learning_rate": 1.894736842105263e-05,
|
| 94 |
+
"loss": 1.6666,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.79182156133829,
|
| 100 |
+
"eval_loss": 0.40839558839797974,
|
| 101 |
+
"eval_macro_f1": 0.741168384879725,
|
| 102 |
+
"eval_runtime": 0.2428,
|
| 103 |
+
"eval_samples_per_second": 1107.725,
|
| 104 |
+
"eval_steps_per_second": 70.005,
|
| 105 |
+
"eval_weighted_f1": 0.7747952835370916,
|
| 106 |
+
"step": 228
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.2913907284768213,
|
| 110 |
+
"grad_norm": 22.404075622558594,
|
| 111 |
+
"learning_rate": 1.8581871345029242e-05,
|
| 112 |
+
"loss": 1.1417,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.622516556291391,
|
| 117 |
+
"grad_norm": 39.31353759765625,
|
| 118 |
+
"learning_rate": 1.821637426900585e-05,
|
| 119 |
+
"loss": 1.2017,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.9536423841059603,
|
| 124 |
+
"grad_norm": 36.49964904785156,
|
| 125 |
+
"learning_rate": 1.785087719298246e-05,
|
| 126 |
+
"loss": 1.1542,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8252788104089219,
|
| 132 |
+
"eval_loss": 0.4079132080078125,
|
| 133 |
+
"eval_macro_f1": 0.7945730766106102,
|
| 134 |
+
"eval_runtime": 0.2474,
|
| 135 |
+
"eval_samples_per_second": 1087.318,
|
| 136 |
+
"eval_steps_per_second": 68.715,
|
| 137 |
+
"eval_weighted_f1": 0.8178976243997124,
|
| 138 |
+
"step": 304
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.2781456953642385,
|
| 142 |
+
"grad_norm": 90.05461120605469,
|
| 143 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 144 |
+
"loss": 0.7798,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.609271523178808,
|
| 149 |
+
"grad_norm": 53.584075927734375,
|
| 150 |
+
"learning_rate": 1.7119883040935675e-05,
|
| 151 |
+
"loss": 1.0223,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.940397350993377,
|
| 156 |
+
"grad_norm": 128.6397705078125,
|
| 157 |
+
"learning_rate": 1.6754385964912282e-05,
|
| 158 |
+
"loss": 0.7932,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.828996282527881,
|
| 164 |
+
"eval_loss": 0.4843341112136841,
|
| 165 |
+
"eval_macro_f1": 0.8068976279650437,
|
| 166 |
+
"eval_runtime": 0.2331,
|
| 167 |
+
"eval_samples_per_second": 1154.141,
|
| 168 |
+
"eval_steps_per_second": 72.938,
|
| 169 |
+
"eval_weighted_f1": 0.8260821742338805,
|
| 170 |
+
"step": 380
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 5.264900662251655,
|
| 174 |
+
"grad_norm": 32.21690368652344,
|
| 175 |
+
"learning_rate": 1.638888888888889e-05,
|
| 176 |
+
"loss": 0.7131,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.596026490066225,
|
| 181 |
+
"grad_norm": 67.67301177978516,
|
| 182 |
+
"learning_rate": 1.60233918128655e-05,
|
| 183 |
+
"loss": 0.698,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.927152317880795,
|
| 188 |
+
"grad_norm": 88.33025360107422,
|
| 189 |
+
"learning_rate": 1.5657894736842107e-05,
|
| 190 |
+
"loss": 0.6571,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 6.0,
|
| 195 |
+
"eval_accuracy": 0.8401486988847584,
|
| 196 |
+
"eval_loss": 0.5245633721351624,
|
| 197 |
+
"eval_macro_f1": 0.8209913799774053,
|
| 198 |
+
"eval_runtime": 0.2399,
|
| 199 |
+
"eval_samples_per_second": 1121.472,
|
| 200 |
+
"eval_steps_per_second": 70.874,
|
| 201 |
+
"eval_weighted_f1": 0.8381894276328701,
|
| 202 |
+
"step": 456
|
| 203 |
+
}
|
| 204 |
+
],
|
| 205 |
+
"logging_steps": 25,
|
| 206 |
+
"max_steps": 1520,
|
| 207 |
+
"num_input_tokens_seen": 0,
|
| 208 |
+
"num_train_epochs": 20,
|
| 209 |
+
"save_steps": 500,
|
| 210 |
+
"stateful_callbacks": {
|
| 211 |
+
"EarlyStoppingCallback": {
|
| 212 |
+
"args": {
|
| 213 |
+
"early_stopping_patience": 3,
|
| 214 |
+
"early_stopping_threshold": 0.0
|
| 215 |
+
},
|
| 216 |
+
"attributes": {
|
| 217 |
+
"early_stopping_patience_counter": 0
|
| 218 |
+
}
|
| 219 |
+
},
|
| 220 |
+
"TrainerControl": {
|
| 221 |
+
"args": {
|
| 222 |
+
"should_epoch_stop": false,
|
| 223 |
+
"should_evaluate": false,
|
| 224 |
+
"should_log": false,
|
| 225 |
+
"should_save": true,
|
| 226 |
+
"should_training_stop": false
|
| 227 |
+
},
|
| 228 |
+
"attributes": {}
|
| 229 |
+
}
|
| 230 |
+
},
|
| 231 |
+
"total_flos": 0.0,
|
| 232 |
+
"train_batch_size": 8,
|
| 233 |
+
"trial_name": null,
|
| 234 |
+
"trial_params": null
|
| 235 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-684/trainer_state.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 456,
|
| 3 |
+
"best_metric": 0.8209913799774053,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-01-s44-20260827T174337Z-fe731097/trainer/checkpoint-456",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 684,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33112582781456956,
|
| 14 |
+
"grad_norm": 29.215723037719727,
|
| 15 |
+
"learning_rate": 3.157894736842105e-06,
|
| 16 |
+
"loss": 2.6441,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6622516556291391,
|
| 21 |
+
"grad_norm": 10.825380325317383,
|
| 22 |
+
"learning_rate": 6.447368421052632e-06,
|
| 23 |
+
"loss": 2.5416,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9933774834437086,
|
| 28 |
+
"grad_norm": 12.073180198669434,
|
| 29 |
+
"learning_rate": 9.736842105263159e-06,
|
| 30 |
+
"loss": 2.5375,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6468401486988847,
|
| 36 |
+
"eval_loss": 0.6156554222106934,
|
| 37 |
+
"eval_macro_f1": 0.3927765237020316,
|
| 38 |
+
"eval_runtime": 0.2515,
|
| 39 |
+
"eval_samples_per_second": 1069.733,
|
| 40 |
+
"eval_steps_per_second": 67.604,
|
| 41 |
+
"eval_weighted_f1": 0.5081272499937063,
|
| 42 |
+
"step": 76
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.3178807947019868,
|
| 46 |
+
"grad_norm": 31.75727081298828,
|
| 47 |
+
"learning_rate": 1.3026315789473684e-05,
|
| 48 |
+
"loss": 2.2574,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6490066225165563,
|
| 53 |
+
"grad_norm": 29.24384307861328,
|
| 54 |
+
"learning_rate": 1.6315789473684213e-05,
|
| 55 |
+
"loss": 2.0182,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.980132450331126,
|
| 60 |
+
"grad_norm": 45.39990997314453,
|
| 61 |
+
"learning_rate": 1.960526315789474e-05,
|
| 62 |
+
"loss": 1.9235,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.7992565055762082,
|
| 68 |
+
"eval_loss": 0.39099982380867004,
|
| 69 |
+
"eval_macro_f1": 0.773314606741573,
|
| 70 |
+
"eval_runtime": 0.2287,
|
| 71 |
+
"eval_samples_per_second": 1176.414,
|
| 72 |
+
"eval_steps_per_second": 74.346,
|
| 73 |
+
"eval_weighted_f1": 0.7958355958397728,
|
| 74 |
+
"step": 152
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.304635761589404,
|
| 78 |
+
"grad_norm": 58.309898376464844,
|
| 79 |
+
"learning_rate": 1.9678362573099416e-05,
|
| 80 |
+
"loss": 1.5733,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.6357615894039736,
|
| 85 |
+
"grad_norm": 35.05423355102539,
|
| 86 |
+
"learning_rate": 1.9312865497076027e-05,
|
| 87 |
+
"loss": 1.5295,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.966887417218543,
|
| 92 |
+
"grad_norm": 15.614839553833008,
|
| 93 |
+
"learning_rate": 1.894736842105263e-05,
|
| 94 |
+
"loss": 1.6666,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.79182156133829,
|
| 100 |
+
"eval_loss": 0.40839558839797974,
|
| 101 |
+
"eval_macro_f1": 0.741168384879725,
|
| 102 |
+
"eval_runtime": 0.2428,
|
| 103 |
+
"eval_samples_per_second": 1107.725,
|
| 104 |
+
"eval_steps_per_second": 70.005,
|
| 105 |
+
"eval_weighted_f1": 0.7747952835370916,
|
| 106 |
+
"step": 228
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.2913907284768213,
|
| 110 |
+
"grad_norm": 22.404075622558594,
|
| 111 |
+
"learning_rate": 1.8581871345029242e-05,
|
| 112 |
+
"loss": 1.1417,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.622516556291391,
|
| 117 |
+
"grad_norm": 39.31353759765625,
|
| 118 |
+
"learning_rate": 1.821637426900585e-05,
|
| 119 |
+
"loss": 1.2017,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.9536423841059603,
|
| 124 |
+
"grad_norm": 36.49964904785156,
|
| 125 |
+
"learning_rate": 1.785087719298246e-05,
|
| 126 |
+
"loss": 1.1542,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8252788104089219,
|
| 132 |
+
"eval_loss": 0.4079132080078125,
|
| 133 |
+
"eval_macro_f1": 0.7945730766106102,
|
| 134 |
+
"eval_runtime": 0.2474,
|
| 135 |
+
"eval_samples_per_second": 1087.318,
|
| 136 |
+
"eval_steps_per_second": 68.715,
|
| 137 |
+
"eval_weighted_f1": 0.8178976243997124,
|
| 138 |
+
"step": 304
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.2781456953642385,
|
| 142 |
+
"grad_norm": 90.05461120605469,
|
| 143 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 144 |
+
"loss": 0.7798,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.609271523178808,
|
| 149 |
+
"grad_norm": 53.584075927734375,
|
| 150 |
+
"learning_rate": 1.7119883040935675e-05,
|
| 151 |
+
"loss": 1.0223,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.940397350993377,
|
| 156 |
+
"grad_norm": 128.6397705078125,
|
| 157 |
+
"learning_rate": 1.6754385964912282e-05,
|
| 158 |
+
"loss": 0.7932,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.828996282527881,
|
| 164 |
+
"eval_loss": 0.4843341112136841,
|
| 165 |
+
"eval_macro_f1": 0.8068976279650437,
|
| 166 |
+
"eval_runtime": 0.2331,
|
| 167 |
+
"eval_samples_per_second": 1154.141,
|
| 168 |
+
"eval_steps_per_second": 72.938,
|
| 169 |
+
"eval_weighted_f1": 0.8260821742338805,
|
| 170 |
+
"step": 380
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 5.264900662251655,
|
| 174 |
+
"grad_norm": 32.21690368652344,
|
| 175 |
+
"learning_rate": 1.638888888888889e-05,
|
| 176 |
+
"loss": 0.7131,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.596026490066225,
|
| 181 |
+
"grad_norm": 67.67301177978516,
|
| 182 |
+
"learning_rate": 1.60233918128655e-05,
|
| 183 |
+
"loss": 0.698,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.927152317880795,
|
| 188 |
+
"grad_norm": 88.33025360107422,
|
| 189 |
+
"learning_rate": 1.5657894736842107e-05,
|
| 190 |
+
"loss": 0.6571,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 6.0,
|
| 195 |
+
"eval_accuracy": 0.8401486988847584,
|
| 196 |
+
"eval_loss": 0.5245633721351624,
|
| 197 |
+
"eval_macro_f1": 0.8209913799774053,
|
| 198 |
+
"eval_runtime": 0.2399,
|
| 199 |
+
"eval_samples_per_second": 1121.472,
|
| 200 |
+
"eval_steps_per_second": 70.874,
|
| 201 |
+
"eval_weighted_f1": 0.8381894276328701,
|
| 202 |
+
"step": 456
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 6.251655629139073,
|
| 206 |
+
"grad_norm": 46.219207763671875,
|
| 207 |
+
"learning_rate": 1.5292397660818715e-05,
|
| 208 |
+
"loss": 0.4157,
|
| 209 |
+
"step": 475
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.582781456953643,
|
| 213 |
+
"grad_norm": 35.7946662902832,
|
| 214 |
+
"learning_rate": 1.4926900584795322e-05,
|
| 215 |
+
"loss": 0.4383,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 6.913907284768212,
|
| 220 |
+
"grad_norm": 171.66189575195312,
|
| 221 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 222 |
+
"loss": 0.6758,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 7.0,
|
| 227 |
+
"eval_accuracy": 0.8327137546468402,
|
| 228 |
+
"eval_loss": 0.725592315196991,
|
| 229 |
+
"eval_macro_f1": 0.7960919733849912,
|
| 230 |
+
"eval_runtime": 0.2439,
|
| 231 |
+
"eval_samples_per_second": 1103.085,
|
| 232 |
+
"eval_steps_per_second": 69.712,
|
| 233 |
+
"eval_weighted_f1": 0.8214702253120619,
|
| 234 |
+
"step": 532
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 7.23841059602649,
|
| 238 |
+
"grad_norm": 5.667948246002197,
|
| 239 |
+
"learning_rate": 1.4195906432748539e-05,
|
| 240 |
+
"loss": 0.3602,
|
| 241 |
+
"step": 550
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.569536423841059,
|
| 245 |
+
"grad_norm": 68.8425521850586,
|
| 246 |
+
"learning_rate": 1.3830409356725148e-05,
|
| 247 |
+
"loss": 0.3528,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 7.900662251655629,
|
| 252 |
+
"grad_norm": 174.67527770996094,
|
| 253 |
+
"learning_rate": 1.3464912280701757e-05,
|
| 254 |
+
"loss": 0.4223,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 8.0,
|
| 259 |
+
"eval_accuracy": 0.8364312267657993,
|
| 260 |
+
"eval_loss": 0.7767894268035889,
|
| 261 |
+
"eval_macro_f1": 0.8095881595881596,
|
| 262 |
+
"eval_runtime": 0.2481,
|
| 263 |
+
"eval_samples_per_second": 1084.114,
|
| 264 |
+
"eval_steps_per_second": 68.513,
|
| 265 |
+
"eval_weighted_f1": 0.8305842220340361,
|
| 266 |
+
"step": 608
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 8.225165562913908,
|
| 270 |
+
"grad_norm": 20.89458656311035,
|
| 271 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 272 |
+
"loss": 0.2944,
|
| 273 |
+
"step": 625
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.556291390728477,
|
| 277 |
+
"grad_norm": 147.40733337402344,
|
| 278 |
+
"learning_rate": 1.2733918128654971e-05,
|
| 279 |
+
"loss": 0.3416,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 8.887417218543046,
|
| 284 |
+
"grad_norm": 60.07394027709961,
|
| 285 |
+
"learning_rate": 1.236842105263158e-05,
|
| 286 |
+
"loss": 0.3217,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 9.0,
|
| 291 |
+
"eval_accuracy": 0.8141263940520446,
|
| 292 |
+
"eval_loss": 0.8457854390144348,
|
| 293 |
+
"eval_macro_f1": 0.7955927051671732,
|
| 294 |
+
"eval_runtime": 0.2554,
|
| 295 |
+
"eval_samples_per_second": 1053.422,
|
| 296 |
+
"eval_steps_per_second": 66.573,
|
| 297 |
+
"eval_weighted_f1": 0.8136687721042699,
|
| 298 |
+
"step": 684
|
| 299 |
+
}
|
| 300 |
+
],
|
| 301 |
+
"logging_steps": 25,
|
| 302 |
+
"max_steps": 1520,
|
| 303 |
+
"num_input_tokens_seen": 0,
|
| 304 |
+
"num_train_epochs": 20,
|
| 305 |
+
"save_steps": 500,
|
| 306 |
+
"stateful_callbacks": {
|
| 307 |
+
"EarlyStoppingCallback": {
|
| 308 |
+
"args": {
|
| 309 |
+
"early_stopping_patience": 3,
|
| 310 |
+
"early_stopping_threshold": 0.0
|
| 311 |
+
},
|
| 312 |
+
"attributes": {
|
| 313 |
+
"early_stopping_patience_counter": 3
|
| 314 |
+
}
|
| 315 |
+
},
|
| 316 |
+
"TrainerControl": {
|
| 317 |
+
"args": {
|
| 318 |
+
"should_epoch_stop": false,
|
| 319 |
+
"should_evaluate": false,
|
| 320 |
+
"should_log": false,
|
| 321 |
+
"should_save": true,
|
| 322 |
+
"should_training_stop": true
|
| 323 |
+
},
|
| 324 |
+
"attributes": {}
|
| 325 |
+
}
|
| 326 |
+
},
|
| 327 |
+
"total_flos": 0.0,
|
| 328 |
+
"train_batch_size": 8,
|
| 329 |
+
"trial_name": null,
|
| 330 |
+
"trial_params": null
|
| 331 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570/trainer_state.json
ADDED
|
@@ -0,0 +1,263 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 570,
|
| 3 |
+
"best_metric": 0.8334779123521725,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 570,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2638522427440633,
|
| 14 |
+
"grad_norm": 28.184856414794922,
|
| 15 |
+
"learning_rate": 2.5263157894736844e-06,
|
| 16 |
+
"loss": 2.582,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5277044854881267,
|
| 21 |
+
"grad_norm": 10.229133605957031,
|
| 22 |
+
"learning_rate": 5.157894736842106e-06,
|
| 23 |
+
"loss": 2.3927,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.7915567282321899,
|
| 28 |
+
"grad_norm": 11.444217681884766,
|
| 29 |
+
"learning_rate": 7.789473684210526e-06,
|
| 30 |
+
"loss": 2.4715,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.712166172106825,
|
| 36 |
+
"eval_loss": 0.5447045564651489,
|
| 37 |
+
"eval_macro_f1": 0.4257531840140536,
|
| 38 |
+
"eval_runtime": 0.3404,
|
| 39 |
+
"eval_samples_per_second": 989.962,
|
| 40 |
+
"eval_steps_per_second": 64.627,
|
| 41 |
+
"eval_weighted_f1": 0.5954348282202753,
|
| 42 |
+
"step": 95
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.0527704485488127,
|
| 46 |
+
"grad_norm": 20.83991050720215,
|
| 47 |
+
"learning_rate": 1.0421052631578948e-05,
|
| 48 |
+
"loss": 2.389,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.316622691292876,
|
| 53 |
+
"grad_norm": 24.766983032226562,
|
| 54 |
+
"learning_rate": 1.305263157894737e-05,
|
| 55 |
+
"loss": 2.1262,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.5804749340369393,
|
| 60 |
+
"grad_norm": 45.65266036987305,
|
| 61 |
+
"learning_rate": 1.568421052631579e-05,
|
| 62 |
+
"loss": 1.9493,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.8443271767810026,
|
| 67 |
+
"grad_norm": 33.494571685791016,
|
| 68 |
+
"learning_rate": 1.831578947368421e-05,
|
| 69 |
+
"loss": 1.7713,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8160237388724035,
|
| 75 |
+
"eval_loss": 0.39242851734161377,
|
| 76 |
+
"eval_macro_f1": 0.7945283612618992,
|
| 77 |
+
"eval_runtime": 0.3138,
|
| 78 |
+
"eval_samples_per_second": 1073.948,
|
| 79 |
+
"eval_steps_per_second": 70.109,
|
| 80 |
+
"eval_weighted_f1": 0.8223343084461298,
|
| 81 |
+
"step": 190
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.1055408970976255,
|
| 85 |
+
"grad_norm": 74.32815551757812,
|
| 86 |
+
"learning_rate": 1.9894736842105265e-05,
|
| 87 |
+
"loss": 1.6568,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.3693931398416885,
|
| 92 |
+
"grad_norm": 23.232336044311523,
|
| 93 |
+
"learning_rate": 1.9602339181286552e-05,
|
| 94 |
+
"loss": 1.7234,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.633245382585752,
|
| 99 |
+
"grad_norm": 20.401756286621094,
|
| 100 |
+
"learning_rate": 1.930994152046784e-05,
|
| 101 |
+
"loss": 1.4142,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.897097625329815,
|
| 106 |
+
"grad_norm": 17.479534149169922,
|
| 107 |
+
"learning_rate": 1.9017543859649123e-05,
|
| 108 |
+
"loss": 1.3028,
|
| 109 |
+
"step": 275
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 3.0,
|
| 113 |
+
"eval_accuracy": 0.8308605341246291,
|
| 114 |
+
"eval_loss": 0.4081993103027344,
|
| 115 |
+
"eval_macro_f1": 0.8122892908445956,
|
| 116 |
+
"eval_runtime": 0.2841,
|
| 117 |
+
"eval_samples_per_second": 1186.245,
|
| 118 |
+
"eval_steps_per_second": 77.44,
|
| 119 |
+
"eval_weighted_f1": 0.8369925484152063,
|
| 120 |
+
"step": 285
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.158311345646438,
|
| 124 |
+
"grad_norm": 49.40342330932617,
|
| 125 |
+
"learning_rate": 1.872514619883041e-05,
|
| 126 |
+
"loss": 1.2521,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.4221635883905015,
|
| 131 |
+
"grad_norm": 54.03762435913086,
|
| 132 |
+
"learning_rate": 1.8432748538011698e-05,
|
| 133 |
+
"loss": 1.0725,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.6860158311345645,
|
| 138 |
+
"grad_norm": 49.46086120605469,
|
| 139 |
+
"learning_rate": 1.8140350877192985e-05,
|
| 140 |
+
"loss": 0.9804,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.949868073878628,
|
| 145 |
+
"grad_norm": 40.94560241699219,
|
| 146 |
+
"learning_rate": 1.784795321637427e-05,
|
| 147 |
+
"loss": 1.1848,
|
| 148 |
+
"step": 375
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 4.0,
|
| 152 |
+
"eval_accuracy": 0.8397626112759644,
|
| 153 |
+
"eval_loss": 0.4325000047683716,
|
| 154 |
+
"eval_macro_f1": 0.7805989583333333,
|
| 155 |
+
"eval_runtime": 0.2986,
|
| 156 |
+
"eval_samples_per_second": 1128.773,
|
| 157 |
+
"eval_steps_per_second": 73.688,
|
| 158 |
+
"eval_weighted_f1": 0.8282679587042532,
|
| 159 |
+
"step": 380
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.211081794195251,
|
| 163 |
+
"grad_norm": 61.03863525390625,
|
| 164 |
+
"learning_rate": 1.7555555555555556e-05,
|
| 165 |
+
"loss": 0.7893,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.474934036939314,
|
| 170 |
+
"grad_norm": 63.03089141845703,
|
| 171 |
+
"learning_rate": 1.7263157894736843e-05,
|
| 172 |
+
"loss": 0.7941,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 4.738786279683377,
|
| 177 |
+
"grad_norm": 141.7689666748047,
|
| 178 |
+
"learning_rate": 1.697076023391813e-05,
|
| 179 |
+
"loss": 0.8326,
|
| 180 |
+
"step": 450
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.0,
|
| 184 |
+
"grad_norm": 24.94400405883789,
|
| 185 |
+
"learning_rate": 1.6678362573099414e-05,
|
| 186 |
+
"loss": 0.8077,
|
| 187 |
+
"step": 475
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 5.0,
|
| 191 |
+
"eval_accuracy": 0.8545994065281899,
|
| 192 |
+
"eval_loss": 0.40618810057640076,
|
| 193 |
+
"eval_macro_f1": 0.8136335421251623,
|
| 194 |
+
"eval_runtime": 0.2885,
|
| 195 |
+
"eval_samples_per_second": 1168.106,
|
| 196 |
+
"eval_steps_per_second": 76.256,
|
| 197 |
+
"eval_weighted_f1": 0.8501916869405224,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.263852242744063,
|
| 202 |
+
"grad_norm": 49.5386962890625,
|
| 203 |
+
"learning_rate": 1.6385964912280702e-05,
|
| 204 |
+
"loss": 0.5325,
|
| 205 |
+
"step": 500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 5.527704485488127,
|
| 209 |
+
"grad_norm": 55.31298828125,
|
| 210 |
+
"learning_rate": 1.609356725146199e-05,
|
| 211 |
+
"loss": 0.6079,
|
| 212 |
+
"step": 525
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 5.79155672823219,
|
| 216 |
+
"grad_norm": 15.50874137878418,
|
| 217 |
+
"learning_rate": 1.5801169590643276e-05,
|
| 218 |
+
"loss": 0.6838,
|
| 219 |
+
"step": 550
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 6.0,
|
| 223 |
+
"eval_accuracy": 0.8664688427299704,
|
| 224 |
+
"eval_loss": 0.43957212567329407,
|
| 225 |
+
"eval_macro_f1": 0.8334779123521725,
|
| 226 |
+
"eval_runtime": 0.291,
|
| 227 |
+
"eval_samples_per_second": 1158.228,
|
| 228 |
+
"eval_steps_per_second": 75.611,
|
| 229 |
+
"eval_weighted_f1": 0.8644893869073026,
|
| 230 |
+
"step": 570
|
| 231 |
+
}
|
| 232 |
+
],
|
| 233 |
+
"logging_steps": 25,
|
| 234 |
+
"max_steps": 1900,
|
| 235 |
+
"num_input_tokens_seen": 0,
|
| 236 |
+
"num_train_epochs": 20,
|
| 237 |
+
"save_steps": 500,
|
| 238 |
+
"stateful_callbacks": {
|
| 239 |
+
"EarlyStoppingCallback": {
|
| 240 |
+
"args": {
|
| 241 |
+
"early_stopping_patience": 3,
|
| 242 |
+
"early_stopping_threshold": 0.0
|
| 243 |
+
},
|
| 244 |
+
"attributes": {
|
| 245 |
+
"early_stopping_patience_counter": 0
|
| 246 |
+
}
|
| 247 |
+
},
|
| 248 |
+
"TrainerControl": {
|
| 249 |
+
"args": {
|
| 250 |
+
"should_epoch_stop": false,
|
| 251 |
+
"should_evaluate": false,
|
| 252 |
+
"should_log": false,
|
| 253 |
+
"should_save": true,
|
| 254 |
+
"should_training_stop": false
|
| 255 |
+
},
|
| 256 |
+
"attributes": {}
|
| 257 |
+
}
|
| 258 |
+
},
|
| 259 |
+
"total_flos": 0.0,
|
| 260 |
+
"train_batch_size": 8,
|
| 261 |
+
"trial_name": null,
|
| 262 |
+
"trial_params": null
|
| 263 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-855/trainer_state.json
ADDED
|
@@ -0,0 +1,380 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 570,
|
| 3 |
+
"best_metric": 0.8334779123521725,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-02-s44-20260827T174508Z-fe731097/trainer/checkpoint-570",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 855,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2638522427440633,
|
| 14 |
+
"grad_norm": 28.184856414794922,
|
| 15 |
+
"learning_rate": 2.5263157894736844e-06,
|
| 16 |
+
"loss": 2.582,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5277044854881267,
|
| 21 |
+
"grad_norm": 10.229133605957031,
|
| 22 |
+
"learning_rate": 5.157894736842106e-06,
|
| 23 |
+
"loss": 2.3927,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.7915567282321899,
|
| 28 |
+
"grad_norm": 11.444217681884766,
|
| 29 |
+
"learning_rate": 7.789473684210526e-06,
|
| 30 |
+
"loss": 2.4715,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.712166172106825,
|
| 36 |
+
"eval_loss": 0.5447045564651489,
|
| 37 |
+
"eval_macro_f1": 0.4257531840140536,
|
| 38 |
+
"eval_runtime": 0.3404,
|
| 39 |
+
"eval_samples_per_second": 989.962,
|
| 40 |
+
"eval_steps_per_second": 64.627,
|
| 41 |
+
"eval_weighted_f1": 0.5954348282202753,
|
| 42 |
+
"step": 95
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.0527704485488127,
|
| 46 |
+
"grad_norm": 20.83991050720215,
|
| 47 |
+
"learning_rate": 1.0421052631578948e-05,
|
| 48 |
+
"loss": 2.389,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.316622691292876,
|
| 53 |
+
"grad_norm": 24.766983032226562,
|
| 54 |
+
"learning_rate": 1.305263157894737e-05,
|
| 55 |
+
"loss": 2.1262,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.5804749340369393,
|
| 60 |
+
"grad_norm": 45.65266036987305,
|
| 61 |
+
"learning_rate": 1.568421052631579e-05,
|
| 62 |
+
"loss": 1.9493,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.8443271767810026,
|
| 67 |
+
"grad_norm": 33.494571685791016,
|
| 68 |
+
"learning_rate": 1.831578947368421e-05,
|
| 69 |
+
"loss": 1.7713,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8160237388724035,
|
| 75 |
+
"eval_loss": 0.39242851734161377,
|
| 76 |
+
"eval_macro_f1": 0.7945283612618992,
|
| 77 |
+
"eval_runtime": 0.3138,
|
| 78 |
+
"eval_samples_per_second": 1073.948,
|
| 79 |
+
"eval_steps_per_second": 70.109,
|
| 80 |
+
"eval_weighted_f1": 0.8223343084461298,
|
| 81 |
+
"step": 190
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.1055408970976255,
|
| 85 |
+
"grad_norm": 74.32815551757812,
|
| 86 |
+
"learning_rate": 1.9894736842105265e-05,
|
| 87 |
+
"loss": 1.6568,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.3693931398416885,
|
| 92 |
+
"grad_norm": 23.232336044311523,
|
| 93 |
+
"learning_rate": 1.9602339181286552e-05,
|
| 94 |
+
"loss": 1.7234,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.633245382585752,
|
| 99 |
+
"grad_norm": 20.401756286621094,
|
| 100 |
+
"learning_rate": 1.930994152046784e-05,
|
| 101 |
+
"loss": 1.4142,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.897097625329815,
|
| 106 |
+
"grad_norm": 17.479534149169922,
|
| 107 |
+
"learning_rate": 1.9017543859649123e-05,
|
| 108 |
+
"loss": 1.3028,
|
| 109 |
+
"step": 275
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 3.0,
|
| 113 |
+
"eval_accuracy": 0.8308605341246291,
|
| 114 |
+
"eval_loss": 0.4081993103027344,
|
| 115 |
+
"eval_macro_f1": 0.8122892908445956,
|
| 116 |
+
"eval_runtime": 0.2841,
|
| 117 |
+
"eval_samples_per_second": 1186.245,
|
| 118 |
+
"eval_steps_per_second": 77.44,
|
| 119 |
+
"eval_weighted_f1": 0.8369925484152063,
|
| 120 |
+
"step": 285
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.158311345646438,
|
| 124 |
+
"grad_norm": 49.40342330932617,
|
| 125 |
+
"learning_rate": 1.872514619883041e-05,
|
| 126 |
+
"loss": 1.2521,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.4221635883905015,
|
| 131 |
+
"grad_norm": 54.03762435913086,
|
| 132 |
+
"learning_rate": 1.8432748538011698e-05,
|
| 133 |
+
"loss": 1.0725,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.6860158311345645,
|
| 138 |
+
"grad_norm": 49.46086120605469,
|
| 139 |
+
"learning_rate": 1.8140350877192985e-05,
|
| 140 |
+
"loss": 0.9804,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.949868073878628,
|
| 145 |
+
"grad_norm": 40.94560241699219,
|
| 146 |
+
"learning_rate": 1.784795321637427e-05,
|
| 147 |
+
"loss": 1.1848,
|
| 148 |
+
"step": 375
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 4.0,
|
| 152 |
+
"eval_accuracy": 0.8397626112759644,
|
| 153 |
+
"eval_loss": 0.4325000047683716,
|
| 154 |
+
"eval_macro_f1": 0.7805989583333333,
|
| 155 |
+
"eval_runtime": 0.2986,
|
| 156 |
+
"eval_samples_per_second": 1128.773,
|
| 157 |
+
"eval_steps_per_second": 73.688,
|
| 158 |
+
"eval_weighted_f1": 0.8282679587042532,
|
| 159 |
+
"step": 380
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.211081794195251,
|
| 163 |
+
"grad_norm": 61.03863525390625,
|
| 164 |
+
"learning_rate": 1.7555555555555556e-05,
|
| 165 |
+
"loss": 0.7893,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.474934036939314,
|
| 170 |
+
"grad_norm": 63.03089141845703,
|
| 171 |
+
"learning_rate": 1.7263157894736843e-05,
|
| 172 |
+
"loss": 0.7941,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 4.738786279683377,
|
| 177 |
+
"grad_norm": 141.7689666748047,
|
| 178 |
+
"learning_rate": 1.697076023391813e-05,
|
| 179 |
+
"loss": 0.8326,
|
| 180 |
+
"step": 450
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.0,
|
| 184 |
+
"grad_norm": 24.94400405883789,
|
| 185 |
+
"learning_rate": 1.6678362573099414e-05,
|
| 186 |
+
"loss": 0.8077,
|
| 187 |
+
"step": 475
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 5.0,
|
| 191 |
+
"eval_accuracy": 0.8545994065281899,
|
| 192 |
+
"eval_loss": 0.40618810057640076,
|
| 193 |
+
"eval_macro_f1": 0.8136335421251623,
|
| 194 |
+
"eval_runtime": 0.2885,
|
| 195 |
+
"eval_samples_per_second": 1168.106,
|
| 196 |
+
"eval_steps_per_second": 76.256,
|
| 197 |
+
"eval_weighted_f1": 0.8501916869405224,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.263852242744063,
|
| 202 |
+
"grad_norm": 49.5386962890625,
|
| 203 |
+
"learning_rate": 1.6385964912280702e-05,
|
| 204 |
+
"loss": 0.5325,
|
| 205 |
+
"step": 500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 5.527704485488127,
|
| 209 |
+
"grad_norm": 55.31298828125,
|
| 210 |
+
"learning_rate": 1.609356725146199e-05,
|
| 211 |
+
"loss": 0.6079,
|
| 212 |
+
"step": 525
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 5.79155672823219,
|
| 216 |
+
"grad_norm": 15.50874137878418,
|
| 217 |
+
"learning_rate": 1.5801169590643276e-05,
|
| 218 |
+
"loss": 0.6838,
|
| 219 |
+
"step": 550
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 6.0,
|
| 223 |
+
"eval_accuracy": 0.8664688427299704,
|
| 224 |
+
"eval_loss": 0.43957212567329407,
|
| 225 |
+
"eval_macro_f1": 0.8334779123521725,
|
| 226 |
+
"eval_runtime": 0.291,
|
| 227 |
+
"eval_samples_per_second": 1158.228,
|
| 228 |
+
"eval_steps_per_second": 75.611,
|
| 229 |
+
"eval_weighted_f1": 0.8644893869073026,
|
| 230 |
+
"step": 570
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 6.052770448548813,
|
| 234 |
+
"grad_norm": 51.36182403564453,
|
| 235 |
+
"learning_rate": 1.5508771929824563e-05,
|
| 236 |
+
"loss": 0.47,
|
| 237 |
+
"step": 575
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 6.316622691292876,
|
| 241 |
+
"grad_norm": 28.888721466064453,
|
| 242 |
+
"learning_rate": 1.5216374269005849e-05,
|
| 243 |
+
"loss": 0.3801,
|
| 244 |
+
"step": 600
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 6.580474934036939,
|
| 248 |
+
"grad_norm": 1.0142254829406738,
|
| 249 |
+
"learning_rate": 1.4923976608187136e-05,
|
| 250 |
+
"loss": 0.3403,
|
| 251 |
+
"step": 625
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 6.844327176781003,
|
| 255 |
+
"grad_norm": 12.541998863220215,
|
| 256 |
+
"learning_rate": 1.4631578947368424e-05,
|
| 257 |
+
"loss": 0.6026,
|
| 258 |
+
"step": 650
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 7.0,
|
| 262 |
+
"eval_accuracy": 0.8605341246290801,
|
| 263 |
+
"eval_loss": 0.5795784592628479,
|
| 264 |
+
"eval_macro_f1": 0.8304266366896846,
|
| 265 |
+
"eval_runtime": 0.2976,
|
| 266 |
+
"eval_samples_per_second": 1132.565,
|
| 267 |
+
"eval_steps_per_second": 73.936,
|
| 268 |
+
"eval_weighted_f1": 0.8603221000661265,
|
| 269 |
+
"step": 665
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 7.105540897097625,
|
| 273 |
+
"grad_norm": 119.6374282836914,
|
| 274 |
+
"learning_rate": 1.4339181286549707e-05,
|
| 275 |
+
"loss": 0.3696,
|
| 276 |
+
"step": 675
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 7.369393139841689,
|
| 280 |
+
"grad_norm": 324.2322082519531,
|
| 281 |
+
"learning_rate": 1.4046783625730995e-05,
|
| 282 |
+
"loss": 0.5652,
|
| 283 |
+
"step": 700
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 7.633245382585752,
|
| 287 |
+
"grad_norm": 45.78263473510742,
|
| 288 |
+
"learning_rate": 1.3754385964912282e-05,
|
| 289 |
+
"loss": 0.4108,
|
| 290 |
+
"step": 725
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 7.897097625329815,
|
| 294 |
+
"grad_norm": 1.5711416006088257,
|
| 295 |
+
"learning_rate": 1.3461988304093567e-05,
|
| 296 |
+
"loss": 0.4606,
|
| 297 |
+
"step": 750
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 8.0,
|
| 301 |
+
"eval_accuracy": 0.857566765578635,
|
| 302 |
+
"eval_loss": 0.8205479383468628,
|
| 303 |
+
"eval_macro_f1": 0.8112044817927171,
|
| 304 |
+
"eval_runtime": 0.3047,
|
| 305 |
+
"eval_samples_per_second": 1105.869,
|
| 306 |
+
"eval_steps_per_second": 72.193,
|
| 307 |
+
"eval_weighted_f1": 0.8503486854682526,
|
| 308 |
+
"step": 760
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 8.158311345646439,
|
| 312 |
+
"grad_norm": 1.0263663530349731,
|
| 313 |
+
"learning_rate": 1.3169590643274855e-05,
|
| 314 |
+
"loss": 0.3653,
|
| 315 |
+
"step": 775
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 8.422163588390502,
|
| 319 |
+
"grad_norm": 333.76715087890625,
|
| 320 |
+
"learning_rate": 1.2877192982456142e-05,
|
| 321 |
+
"loss": 0.2266,
|
| 322 |
+
"step": 800
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 8.686015831134565,
|
| 326 |
+
"grad_norm": 490.29718017578125,
|
| 327 |
+
"learning_rate": 1.258479532163743e-05,
|
| 328 |
+
"loss": 0.4446,
|
| 329 |
+
"step": 825
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 8.949868073878628,
|
| 333 |
+
"grad_norm": 1.7823795080184937,
|
| 334 |
+
"learning_rate": 1.2292397660818713e-05,
|
| 335 |
+
"loss": 0.4719,
|
| 336 |
+
"step": 850
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 9.0,
|
| 340 |
+
"eval_accuracy": 0.8427299703264095,
|
| 341 |
+
"eval_loss": 0.9018192291259766,
|
| 342 |
+
"eval_macro_f1": 0.7837468066301019,
|
| 343 |
+
"eval_runtime": 0.3188,
|
| 344 |
+
"eval_samples_per_second": 1057.249,
|
| 345 |
+
"eval_steps_per_second": 69.019,
|
| 346 |
+
"eval_weighted_f1": 0.83100036390953,
|
| 347 |
+
"step": 855
|
| 348 |
+
}
|
| 349 |
+
],
|
| 350 |
+
"logging_steps": 25,
|
| 351 |
+
"max_steps": 1900,
|
| 352 |
+
"num_input_tokens_seen": 0,
|
| 353 |
+
"num_train_epochs": 20,
|
| 354 |
+
"save_steps": 500,
|
| 355 |
+
"stateful_callbacks": {
|
| 356 |
+
"EarlyStoppingCallback": {
|
| 357 |
+
"args": {
|
| 358 |
+
"early_stopping_patience": 3,
|
| 359 |
+
"early_stopping_threshold": 0.0
|
| 360 |
+
},
|
| 361 |
+
"attributes": {
|
| 362 |
+
"early_stopping_patience_counter": 3
|
| 363 |
+
}
|
| 364 |
+
},
|
| 365 |
+
"TrainerControl": {
|
| 366 |
+
"args": {
|
| 367 |
+
"should_epoch_stop": false,
|
| 368 |
+
"should_evaluate": false,
|
| 369 |
+
"should_log": false,
|
| 370 |
+
"should_save": true,
|
| 371 |
+
"should_training_stop": true
|
| 372 |
+
},
|
| 373 |
+
"attributes": {}
|
| 374 |
+
}
|
| 375 |
+
},
|
| 376 |
+
"total_flos": 0.0,
|
| 377 |
+
"train_batch_size": 8,
|
| 378 |
+
"trial_name": null,
|
| 379 |
+
"trial_params": null
|
| 380 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440/trainer_state.json
ADDED
|
@@ -0,0 +1,217 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 440,
|
| 3 |
+
"best_metric": 0.82947101678809,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 440,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2840909090909091,
|
| 14 |
+
"grad_norm": 11.185952186584473,
|
| 15 |
+
"learning_rate": 2.7272727272727272e-06,
|
| 16 |
+
"loss": 2.5237,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5681818181818182,
|
| 21 |
+
"grad_norm": 7.978830337524414,
|
| 22 |
+
"learning_rate": 5.568181818181818e-06,
|
| 23 |
+
"loss": 2.3503,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.8522727272727273,
|
| 28 |
+
"grad_norm": 16.181005477905273,
|
| 29 |
+
"learning_rate": 8.40909090909091e-06,
|
| 30 |
+
"loss": 2.4642,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.7220447284345048,
|
| 36 |
+
"eval_loss": 0.5378425717353821,
|
| 37 |
+
"eval_macro_f1": 0.41929499072356213,
|
| 38 |
+
"eval_runtime": 0.2678,
|
| 39 |
+
"eval_samples_per_second": 1168.884,
|
| 40 |
+
"eval_steps_per_second": 74.689,
|
| 41 |
+
"eval_weighted_f1": 0.6054994754218853,
|
| 42 |
+
"step": 88
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.1363636363636362,
|
| 46 |
+
"grad_norm": 30.144990921020508,
|
| 47 |
+
"learning_rate": 1.125e-05,
|
| 48 |
+
"loss": 2.2965,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.4204545454545454,
|
| 53 |
+
"grad_norm": 57.358558654785156,
|
| 54 |
+
"learning_rate": 1.4090909090909092e-05,
|
| 55 |
+
"loss": 1.9431,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.7045454545454546,
|
| 60 |
+
"grad_norm": 53.61069107055664,
|
| 61 |
+
"learning_rate": 1.6931818181818182e-05,
|
| 62 |
+
"loss": 1.7463,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.9886363636363638,
|
| 67 |
+
"grad_norm": 21.49080467224121,
|
| 68 |
+
"learning_rate": 1.9772727272727274e-05,
|
| 69 |
+
"loss": 1.8495,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8434504792332268,
|
| 75 |
+
"eval_loss": 0.3734992742538452,
|
| 76 |
+
"eval_macro_f1": 0.7934494229189393,
|
| 77 |
+
"eval_runtime": 0.2708,
|
| 78 |
+
"eval_samples_per_second": 1155.91,
|
| 79 |
+
"eval_steps_per_second": 73.86,
|
| 80 |
+
"eval_weighted_f1": 0.8385802464753416,
|
| 81 |
+
"step": 176
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.2727272727272725,
|
| 85 |
+
"grad_norm": 56.5705451965332,
|
| 86 |
+
"learning_rate": 1.9709595959595963e-05,
|
| 87 |
+
"loss": 1.762,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.5568181818181817,
|
| 92 |
+
"grad_norm": 43.73651123046875,
|
| 93 |
+
"learning_rate": 1.9393939393939395e-05,
|
| 94 |
+
"loss": 1.382,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.840909090909091,
|
| 99 |
+
"grad_norm": 46.1258430480957,
|
| 100 |
+
"learning_rate": 1.907828282828283e-05,
|
| 101 |
+
"loss": 1.4919,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_accuracy": 0.853035143769968,
|
| 107 |
+
"eval_loss": 0.3199264109134674,
|
| 108 |
+
"eval_macro_f1": 0.8291646891314666,
|
| 109 |
+
"eval_runtime": 0.2599,
|
| 110 |
+
"eval_samples_per_second": 1204.251,
|
| 111 |
+
"eval_steps_per_second": 76.949,
|
| 112 |
+
"eval_weighted_f1": 0.8575236053259255,
|
| 113 |
+
"step": 264
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.125,
|
| 117 |
+
"grad_norm": 35.986595153808594,
|
| 118 |
+
"learning_rate": 1.8762626262626265e-05,
|
| 119 |
+
"loss": 1.3461,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.409090909090909,
|
| 124 |
+
"grad_norm": 36.81462860107422,
|
| 125 |
+
"learning_rate": 1.84469696969697e-05,
|
| 126 |
+
"loss": 1.3021,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.6931818181818183,
|
| 131 |
+
"grad_norm": 31.702444076538086,
|
| 132 |
+
"learning_rate": 1.8131313131313132e-05,
|
| 133 |
+
"loss": 0.981,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.9772727272727275,
|
| 138 |
+
"grad_norm": 41.3076171875,
|
| 139 |
+
"learning_rate": 1.7815656565656567e-05,
|
| 140 |
+
"loss": 0.9622,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.8626198083067093,
|
| 146 |
+
"eval_loss": 0.3892499506473541,
|
| 147 |
+
"eval_macro_f1": 0.8187413303166202,
|
| 148 |
+
"eval_runtime": 0.2598,
|
| 149 |
+
"eval_samples_per_second": 1204.942,
|
| 150 |
+
"eval_steps_per_second": 76.993,
|
| 151 |
+
"eval_weighted_f1": 0.8583459305804019,
|
| 152 |
+
"step": 352
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.261363636363637,
|
| 156 |
+
"grad_norm": 18.65777587890625,
|
| 157 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 158 |
+
"loss": 0.7881,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.545454545454545,
|
| 163 |
+
"grad_norm": 55.96082305908203,
|
| 164 |
+
"learning_rate": 1.7184343434343437e-05,
|
| 165 |
+
"loss": 0.756,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.829545454545455,
|
| 170 |
+
"grad_norm": 19.324739456176758,
|
| 171 |
+
"learning_rate": 1.686868686868687e-05,
|
| 172 |
+
"loss": 0.6801,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 5.0,
|
| 177 |
+
"eval_accuracy": 0.8626198083067093,
|
| 178 |
+
"eval_loss": 0.3863547444343567,
|
| 179 |
+
"eval_macro_f1": 0.82947101678809,
|
| 180 |
+
"eval_runtime": 0.2647,
|
| 181 |
+
"eval_samples_per_second": 1182.667,
|
| 182 |
+
"eval_steps_per_second": 75.57,
|
| 183 |
+
"eval_weighted_f1": 0.8628600169409022,
|
| 184 |
+
"step": 440
|
| 185 |
+
}
|
| 186 |
+
],
|
| 187 |
+
"logging_steps": 25,
|
| 188 |
+
"max_steps": 1760,
|
| 189 |
+
"num_input_tokens_seen": 0,
|
| 190 |
+
"num_train_epochs": 20,
|
| 191 |
+
"save_steps": 500,
|
| 192 |
+
"stateful_callbacks": {
|
| 193 |
+
"EarlyStoppingCallback": {
|
| 194 |
+
"args": {
|
| 195 |
+
"early_stopping_patience": 3,
|
| 196 |
+
"early_stopping_threshold": 0.0
|
| 197 |
+
},
|
| 198 |
+
"attributes": {
|
| 199 |
+
"early_stopping_patience_counter": 0
|
| 200 |
+
}
|
| 201 |
+
},
|
| 202 |
+
"TrainerControl": {
|
| 203 |
+
"args": {
|
| 204 |
+
"should_epoch_stop": false,
|
| 205 |
+
"should_evaluate": false,
|
| 206 |
+
"should_log": false,
|
| 207 |
+
"should_save": true,
|
| 208 |
+
"should_training_stop": false
|
| 209 |
+
},
|
| 210 |
+
"attributes": {}
|
| 211 |
+
}
|
| 212 |
+
},
|
| 213 |
+
"total_flos": 0.0,
|
| 214 |
+
"train_batch_size": 8,
|
| 215 |
+
"trial_name": null,
|
| 216 |
+
"trial_params": null
|
| 217 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-704/trainer_state.json
ADDED
|
@@ -0,0 +1,327 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 440,
|
| 3 |
+
"best_metric": 0.82947101678809,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-03-s44-20260827T174658Z-fe731097/trainer/checkpoint-440",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 704,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2840909090909091,
|
| 14 |
+
"grad_norm": 11.185952186584473,
|
| 15 |
+
"learning_rate": 2.7272727272727272e-06,
|
| 16 |
+
"loss": 2.5237,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5681818181818182,
|
| 21 |
+
"grad_norm": 7.978830337524414,
|
| 22 |
+
"learning_rate": 5.568181818181818e-06,
|
| 23 |
+
"loss": 2.3503,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.8522727272727273,
|
| 28 |
+
"grad_norm": 16.181005477905273,
|
| 29 |
+
"learning_rate": 8.40909090909091e-06,
|
| 30 |
+
"loss": 2.4642,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.7220447284345048,
|
| 36 |
+
"eval_loss": 0.5378425717353821,
|
| 37 |
+
"eval_macro_f1": 0.41929499072356213,
|
| 38 |
+
"eval_runtime": 0.2678,
|
| 39 |
+
"eval_samples_per_second": 1168.884,
|
| 40 |
+
"eval_steps_per_second": 74.689,
|
| 41 |
+
"eval_weighted_f1": 0.6054994754218853,
|
| 42 |
+
"step": 88
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.1363636363636362,
|
| 46 |
+
"grad_norm": 30.144990921020508,
|
| 47 |
+
"learning_rate": 1.125e-05,
|
| 48 |
+
"loss": 2.2965,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.4204545454545454,
|
| 53 |
+
"grad_norm": 57.358558654785156,
|
| 54 |
+
"learning_rate": 1.4090909090909092e-05,
|
| 55 |
+
"loss": 1.9431,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.7045454545454546,
|
| 60 |
+
"grad_norm": 53.61069107055664,
|
| 61 |
+
"learning_rate": 1.6931818181818182e-05,
|
| 62 |
+
"loss": 1.7463,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.9886363636363638,
|
| 67 |
+
"grad_norm": 21.49080467224121,
|
| 68 |
+
"learning_rate": 1.9772727272727274e-05,
|
| 69 |
+
"loss": 1.8495,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8434504792332268,
|
| 75 |
+
"eval_loss": 0.3734992742538452,
|
| 76 |
+
"eval_macro_f1": 0.7934494229189393,
|
| 77 |
+
"eval_runtime": 0.2708,
|
| 78 |
+
"eval_samples_per_second": 1155.91,
|
| 79 |
+
"eval_steps_per_second": 73.86,
|
| 80 |
+
"eval_weighted_f1": 0.8385802464753416,
|
| 81 |
+
"step": 176
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.2727272727272725,
|
| 85 |
+
"grad_norm": 56.5705451965332,
|
| 86 |
+
"learning_rate": 1.9709595959595963e-05,
|
| 87 |
+
"loss": 1.762,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.5568181818181817,
|
| 92 |
+
"grad_norm": 43.73651123046875,
|
| 93 |
+
"learning_rate": 1.9393939393939395e-05,
|
| 94 |
+
"loss": 1.382,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.840909090909091,
|
| 99 |
+
"grad_norm": 46.1258430480957,
|
| 100 |
+
"learning_rate": 1.907828282828283e-05,
|
| 101 |
+
"loss": 1.4919,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_accuracy": 0.853035143769968,
|
| 107 |
+
"eval_loss": 0.3199264109134674,
|
| 108 |
+
"eval_macro_f1": 0.8291646891314666,
|
| 109 |
+
"eval_runtime": 0.2599,
|
| 110 |
+
"eval_samples_per_second": 1204.251,
|
| 111 |
+
"eval_steps_per_second": 76.949,
|
| 112 |
+
"eval_weighted_f1": 0.8575236053259255,
|
| 113 |
+
"step": 264
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.125,
|
| 117 |
+
"grad_norm": 35.986595153808594,
|
| 118 |
+
"learning_rate": 1.8762626262626265e-05,
|
| 119 |
+
"loss": 1.3461,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.409090909090909,
|
| 124 |
+
"grad_norm": 36.81462860107422,
|
| 125 |
+
"learning_rate": 1.84469696969697e-05,
|
| 126 |
+
"loss": 1.3021,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.6931818181818183,
|
| 131 |
+
"grad_norm": 31.702444076538086,
|
| 132 |
+
"learning_rate": 1.8131313131313132e-05,
|
| 133 |
+
"loss": 0.981,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.9772727272727275,
|
| 138 |
+
"grad_norm": 41.3076171875,
|
| 139 |
+
"learning_rate": 1.7815656565656567e-05,
|
| 140 |
+
"loss": 0.9622,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.8626198083067093,
|
| 146 |
+
"eval_loss": 0.3892499506473541,
|
| 147 |
+
"eval_macro_f1": 0.8187413303166202,
|
| 148 |
+
"eval_runtime": 0.2598,
|
| 149 |
+
"eval_samples_per_second": 1204.942,
|
| 150 |
+
"eval_steps_per_second": 76.993,
|
| 151 |
+
"eval_weighted_f1": 0.8583459305804019,
|
| 152 |
+
"step": 352
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.261363636363637,
|
| 156 |
+
"grad_norm": 18.65777587890625,
|
| 157 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 158 |
+
"loss": 0.7881,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.545454545454545,
|
| 163 |
+
"grad_norm": 55.96082305908203,
|
| 164 |
+
"learning_rate": 1.7184343434343437e-05,
|
| 165 |
+
"loss": 0.756,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.829545454545455,
|
| 170 |
+
"grad_norm": 19.324739456176758,
|
| 171 |
+
"learning_rate": 1.686868686868687e-05,
|
| 172 |
+
"loss": 0.6801,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 5.0,
|
| 177 |
+
"eval_accuracy": 0.8626198083067093,
|
| 178 |
+
"eval_loss": 0.3863547444343567,
|
| 179 |
+
"eval_macro_f1": 0.82947101678809,
|
| 180 |
+
"eval_runtime": 0.2647,
|
| 181 |
+
"eval_samples_per_second": 1182.667,
|
| 182 |
+
"eval_steps_per_second": 75.57,
|
| 183 |
+
"eval_weighted_f1": 0.8628600169409022,
|
| 184 |
+
"step": 440
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.113636363636363,
|
| 188 |
+
"grad_norm": 65.35336303710938,
|
| 189 |
+
"learning_rate": 1.6553030303030304e-05,
|
| 190 |
+
"loss": 0.5734,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.3977272727272725,
|
| 195 |
+
"grad_norm": 20.20461654663086,
|
| 196 |
+
"learning_rate": 1.623737373737374e-05,
|
| 197 |
+
"loss": 0.4503,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.681818181818182,
|
| 202 |
+
"grad_norm": 22.524442672729492,
|
| 203 |
+
"learning_rate": 1.5921717171717174e-05,
|
| 204 |
+
"loss": 0.6496,
|
| 205 |
+
"step": 500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 5.965909090909091,
|
| 209 |
+
"grad_norm": 40.10174560546875,
|
| 210 |
+
"learning_rate": 1.5606060606060605e-05,
|
| 211 |
+
"loss": 0.5695,
|
| 212 |
+
"step": 525
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 6.0,
|
| 216 |
+
"eval_accuracy": 0.8306709265175719,
|
| 217 |
+
"eval_loss": 0.48975321650505066,
|
| 218 |
+
"eval_macro_f1": 0.7801878917171289,
|
| 219 |
+
"eval_runtime": 0.2699,
|
| 220 |
+
"eval_samples_per_second": 1159.872,
|
| 221 |
+
"eval_steps_per_second": 74.113,
|
| 222 |
+
"eval_weighted_f1": 0.8269688372988727,
|
| 223 |
+
"step": 528
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.25,
|
| 227 |
+
"grad_norm": 66.32112121582031,
|
| 228 |
+
"learning_rate": 1.529040404040404e-05,
|
| 229 |
+
"loss": 0.4683,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 6.534090909090909,
|
| 234 |
+
"grad_norm": 101.5428695678711,
|
| 235 |
+
"learning_rate": 1.4974747474747475e-05,
|
| 236 |
+
"loss": 0.4522,
|
| 237 |
+
"step": 575
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 6.818181818181818,
|
| 241 |
+
"grad_norm": 120.77235412597656,
|
| 242 |
+
"learning_rate": 1.465909090909091e-05,
|
| 243 |
+
"loss": 0.429,
|
| 244 |
+
"step": 600
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 7.0,
|
| 248 |
+
"eval_accuracy": 0.8402555910543131,
|
| 249 |
+
"eval_loss": 0.6184565424919128,
|
| 250 |
+
"eval_macro_f1": 0.780812324929972,
|
| 251 |
+
"eval_runtime": 0.265,
|
| 252 |
+
"eval_samples_per_second": 1181.341,
|
| 253 |
+
"eval_steps_per_second": 75.485,
|
| 254 |
+
"eval_weighted_f1": 0.8315032083120787,
|
| 255 |
+
"step": 616
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 7.1022727272727275,
|
| 259 |
+
"grad_norm": 83.2934799194336,
|
| 260 |
+
"learning_rate": 1.4343434343434344e-05,
|
| 261 |
+
"loss": 0.4202,
|
| 262 |
+
"step": 625
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 7.386363636363637,
|
| 266 |
+
"grad_norm": 69.76300811767578,
|
| 267 |
+
"learning_rate": 1.4027777777777779e-05,
|
| 268 |
+
"loss": 0.4284,
|
| 269 |
+
"step": 650
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 7.670454545454545,
|
| 273 |
+
"grad_norm": 82.83062744140625,
|
| 274 |
+
"learning_rate": 1.3712121212121212e-05,
|
| 275 |
+
"loss": 0.5921,
|
| 276 |
+
"step": 675
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 7.954545454545455,
|
| 280 |
+
"grad_norm": 192.17881774902344,
|
| 281 |
+
"learning_rate": 1.3396464646464649e-05,
|
| 282 |
+
"loss": 0.2212,
|
| 283 |
+
"step": 700
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 8.0,
|
| 287 |
+
"eval_accuracy": 0.8434504792332268,
|
| 288 |
+
"eval_loss": 0.8682672381401062,
|
| 289 |
+
"eval_macro_f1": 0.7967774847950815,
|
| 290 |
+
"eval_runtime": 0.2722,
|
| 291 |
+
"eval_samples_per_second": 1149.86,
|
| 292 |
+
"eval_steps_per_second": 73.473,
|
| 293 |
+
"eval_weighted_f1": 0.8400277929744295,
|
| 294 |
+
"step": 704
|
| 295 |
+
}
|
| 296 |
+
],
|
| 297 |
+
"logging_steps": 25,
|
| 298 |
+
"max_steps": 1760,
|
| 299 |
+
"num_input_tokens_seen": 0,
|
| 300 |
+
"num_train_epochs": 20,
|
| 301 |
+
"save_steps": 500,
|
| 302 |
+
"stateful_callbacks": {
|
| 303 |
+
"EarlyStoppingCallback": {
|
| 304 |
+
"args": {
|
| 305 |
+
"early_stopping_patience": 3,
|
| 306 |
+
"early_stopping_threshold": 0.0
|
| 307 |
+
},
|
| 308 |
+
"attributes": {
|
| 309 |
+
"early_stopping_patience_counter": 3
|
| 310 |
+
}
|
| 311 |
+
},
|
| 312 |
+
"TrainerControl": {
|
| 313 |
+
"args": {
|
| 314 |
+
"should_epoch_stop": false,
|
| 315 |
+
"should_evaluate": false,
|
| 316 |
+
"should_log": false,
|
| 317 |
+
"should_save": true,
|
| 318 |
+
"should_training_stop": true
|
| 319 |
+
},
|
| 320 |
+
"attributes": {}
|
| 321 |
+
}
|
| 322 |
+
},
|
| 323 |
+
"total_flos": 0.0,
|
| 324 |
+
"train_batch_size": 8,
|
| 325 |
+
"trial_name": null,
|
| 326 |
+
"trial_params": null
|
| 327 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-1040/trainer_state.json
ADDED
|
@@ -0,0 +1,473 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 800,
|
| 3 |
+
"best_metric": 0.882421802504033,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800",
|
| 5 |
+
"epoch": 13.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1040,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.3125,
|
| 14 |
+
"grad_norm": 17.846830368041992,
|
| 15 |
+
"learning_rate": 3e-06,
|
| 16 |
+
"loss": 2.5728,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.625,
|
| 21 |
+
"grad_norm": 11.913187980651855,
|
| 22 |
+
"learning_rate": 6.125000000000001e-06,
|
| 23 |
+
"loss": 2.5396,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9375,
|
| 28 |
+
"grad_norm": 9.724447250366211,
|
| 29 |
+
"learning_rate": 9.250000000000001e-06,
|
| 30 |
+
"loss": 2.4403,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.676056338028169,
|
| 36 |
+
"eval_loss": 0.5894998908042908,
|
| 37 |
+
"eval_macro_f1": 0.40336134453781514,
|
| 38 |
+
"eval_runtime": 0.2623,
|
| 39 |
+
"eval_samples_per_second": 1082.721,
|
| 40 |
+
"eval_steps_per_second": 68.623,
|
| 41 |
+
"eval_weighted_f1": 0.5453899869807078,
|
| 42 |
+
"step": 80
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.25,
|
| 46 |
+
"grad_norm": 27.37001609802246,
|
| 47 |
+
"learning_rate": 1.2375000000000001e-05,
|
| 48 |
+
"loss": 2.2358,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.5625,
|
| 53 |
+
"grad_norm": 24.617549896240234,
|
| 54 |
+
"learning_rate": 1.55e-05,
|
| 55 |
+
"loss": 2.0279,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.875,
|
| 60 |
+
"grad_norm": 24.2894344329834,
|
| 61 |
+
"learning_rate": 1.8625000000000002e-05,
|
| 62 |
+
"loss": 1.9649,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8415492957746479,
|
| 68 |
+
"eval_loss": 0.38190901279449463,
|
| 69 |
+
"eval_macro_f1": 0.8129801712153362,
|
| 70 |
+
"eval_runtime": 0.2571,
|
| 71 |
+
"eval_samples_per_second": 1104.709,
|
| 72 |
+
"eval_steps_per_second": 70.017,
|
| 73 |
+
"eval_weighted_f1": 0.8387181212687701,
|
| 74 |
+
"step": 160
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.1875,
|
| 78 |
+
"grad_norm": 86.57966613769531,
|
| 79 |
+
"learning_rate": 1.9805555555555557e-05,
|
| 80 |
+
"loss": 1.6063,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.5,
|
| 85 |
+
"grad_norm": 45.83436965942383,
|
| 86 |
+
"learning_rate": 1.9458333333333333e-05,
|
| 87 |
+
"loss": 1.6571,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.8125,
|
| 92 |
+
"grad_norm": 49.14204406738281,
|
| 93 |
+
"learning_rate": 1.9111111111111113e-05,
|
| 94 |
+
"loss": 1.5296,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8732394366197183,
|
| 100 |
+
"eval_loss": 0.3036462366580963,
|
| 101 |
+
"eval_macro_f1": 0.8536082474226805,
|
| 102 |
+
"eval_runtime": 0.2673,
|
| 103 |
+
"eval_samples_per_second": 1062.583,
|
| 104 |
+
"eval_steps_per_second": 67.347,
|
| 105 |
+
"eval_weighted_f1": 0.8724843908813708,
|
| 106 |
+
"step": 240
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.125,
|
| 110 |
+
"grad_norm": 41.95870590209961,
|
| 111 |
+
"learning_rate": 1.876388888888889e-05,
|
| 112 |
+
"loss": 1.3436,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.4375,
|
| 117 |
+
"grad_norm": 26.975175857543945,
|
| 118 |
+
"learning_rate": 1.8416666666666666e-05,
|
| 119 |
+
"loss": 1.0426,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.75,
|
| 124 |
+
"grad_norm": 36.39384841918945,
|
| 125 |
+
"learning_rate": 1.8069444444444446e-05,
|
| 126 |
+
"loss": 1.2598,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8661971830985915,
|
| 132 |
+
"eval_loss": 0.32074618339538574,
|
| 133 |
+
"eval_macro_f1": 0.8359180198260658,
|
| 134 |
+
"eval_runtime": 0.262,
|
| 135 |
+
"eval_samples_per_second": 1084.045,
|
| 136 |
+
"eval_steps_per_second": 68.707,
|
| 137 |
+
"eval_weighted_f1": 0.8607370061150214,
|
| 138 |
+
"step": 320
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.0625,
|
| 142 |
+
"grad_norm": 34.646976470947266,
|
| 143 |
+
"learning_rate": 1.7722222222222222e-05,
|
| 144 |
+
"loss": 1.139,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.375,
|
| 149 |
+
"grad_norm": 53.85174560546875,
|
| 150 |
+
"learning_rate": 1.7375000000000002e-05,
|
| 151 |
+
"loss": 0.6466,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.6875,
|
| 156 |
+
"grad_norm": 39.1696891784668,
|
| 157 |
+
"learning_rate": 1.702777777777778e-05,
|
| 158 |
+
"loss": 0.9732,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"grad_norm": 67.7001724243164,
|
| 164 |
+
"learning_rate": 1.668055555555556e-05,
|
| 165 |
+
"loss": 1.0013,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 5.0,
|
| 170 |
+
"eval_accuracy": 0.8802816901408451,
|
| 171 |
+
"eval_loss": 0.34587180614471436,
|
| 172 |
+
"eval_macro_f1": 0.857286432160804,
|
| 173 |
+
"eval_runtime": 0.2552,
|
| 174 |
+
"eval_samples_per_second": 1112.745,
|
| 175 |
+
"eval_steps_per_second": 70.526,
|
| 176 |
+
"eval_weighted_f1": 0.8774577110906647,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.3125,
|
| 181 |
+
"grad_norm": 99.6565170288086,
|
| 182 |
+
"learning_rate": 1.6333333333333335e-05,
|
| 183 |
+
"loss": 0.7472,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.625,
|
| 188 |
+
"grad_norm": 11.587423324584961,
|
| 189 |
+
"learning_rate": 1.5986111111111115e-05,
|
| 190 |
+
"loss": 0.7773,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.9375,
|
| 195 |
+
"grad_norm": 27.089900970458984,
|
| 196 |
+
"learning_rate": 1.563888888888889e-05,
|
| 197 |
+
"loss": 0.4622,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 6.0,
|
| 202 |
+
"eval_accuracy": 0.8415492957746479,
|
| 203 |
+
"eval_loss": 0.6065791845321655,
|
| 204 |
+
"eval_macro_f1": 0.7897161661867544,
|
| 205 |
+
"eval_runtime": 0.3906,
|
| 206 |
+
"eval_samples_per_second": 727.002,
|
| 207 |
+
"eval_steps_per_second": 46.078,
|
| 208 |
+
"eval_weighted_f1": 0.8264772510008633,
|
| 209 |
+
"step": 480
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.25,
|
| 213 |
+
"grad_norm": 46.45066833496094,
|
| 214 |
+
"learning_rate": 1.5291666666666668e-05,
|
| 215 |
+
"loss": 0.5752,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 6.5625,
|
| 220 |
+
"grad_norm": 39.581424713134766,
|
| 221 |
+
"learning_rate": 1.4944444444444446e-05,
|
| 222 |
+
"loss": 0.6317,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.875,
|
| 227 |
+
"grad_norm": 22.54334831237793,
|
| 228 |
+
"learning_rate": 1.4597222222222223e-05,
|
| 229 |
+
"loss": 0.3716,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 7.0,
|
| 234 |
+
"eval_accuracy": 0.8556338028169014,
|
| 235 |
+
"eval_loss": 0.5141348242759705,
|
| 236 |
+
"eval_macro_f1": 0.8391157167530225,
|
| 237 |
+
"eval_runtime": 0.2782,
|
| 238 |
+
"eval_samples_per_second": 1020.909,
|
| 239 |
+
"eval_steps_per_second": 64.706,
|
| 240 |
+
"eval_weighted_f1": 0.857267459680362,
|
| 241 |
+
"step": 560
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.1875,
|
| 245 |
+
"grad_norm": 80.10563659667969,
|
| 246 |
+
"learning_rate": 1.425e-05,
|
| 247 |
+
"loss": 0.4154,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 7.5,
|
| 252 |
+
"grad_norm": 34.158931732177734,
|
| 253 |
+
"learning_rate": 1.3902777777777779e-05,
|
| 254 |
+
"loss": 0.4169,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 7.8125,
|
| 259 |
+
"grad_norm": 4.919360637664795,
|
| 260 |
+
"learning_rate": 1.3555555555555557e-05,
|
| 261 |
+
"loss": 0.3483,
|
| 262 |
+
"step": 625
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 8.0,
|
| 266 |
+
"eval_accuracy": 0.8802816901408451,
|
| 267 |
+
"eval_loss": 0.5978385210037231,
|
| 268 |
+
"eval_macro_f1": 0.8582334977683814,
|
| 269 |
+
"eval_runtime": 0.2702,
|
| 270 |
+
"eval_samples_per_second": 1050.961,
|
| 271 |
+
"eval_steps_per_second": 66.61,
|
| 272 |
+
"eval_weighted_f1": 0.8779193838152238,
|
| 273 |
+
"step": 640
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.125,
|
| 277 |
+
"grad_norm": 7.283858299255371,
|
| 278 |
+
"learning_rate": 1.3208333333333334e-05,
|
| 279 |
+
"loss": 0.3996,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 8.4375,
|
| 284 |
+
"grad_norm": 7.524752140045166,
|
| 285 |
+
"learning_rate": 1.2861111111111114e-05,
|
| 286 |
+
"loss": 0.3507,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 8.75,
|
| 291 |
+
"grad_norm": 83.3502197265625,
|
| 292 |
+
"learning_rate": 1.251388888888889e-05,
|
| 293 |
+
"loss": 0.2851,
|
| 294 |
+
"step": 700
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 9.0,
|
| 298 |
+
"eval_accuracy": 0.8767605633802817,
|
| 299 |
+
"eval_loss": 0.6831046342849731,
|
| 300 |
+
"eval_macro_f1": 0.855470737913486,
|
| 301 |
+
"eval_runtime": 0.2725,
|
| 302 |
+
"eval_samples_per_second": 1042.115,
|
| 303 |
+
"eval_steps_per_second": 66.05,
|
| 304 |
+
"eval_weighted_f1": 0.8750026878830235,
|
| 305 |
+
"step": 720
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 9.0625,
|
| 309 |
+
"grad_norm": 0.4345070421695709,
|
| 310 |
+
"learning_rate": 1.2166666666666667e-05,
|
| 311 |
+
"loss": 0.3616,
|
| 312 |
+
"step": 725
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 9.375,
|
| 316 |
+
"grad_norm": 0.09883200377225876,
|
| 317 |
+
"learning_rate": 1.1819444444444446e-05,
|
| 318 |
+
"loss": 0.2566,
|
| 319 |
+
"step": 750
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 9.6875,
|
| 323 |
+
"grad_norm": 56.23919677734375,
|
| 324 |
+
"learning_rate": 1.1472222222222223e-05,
|
| 325 |
+
"loss": 0.3902,
|
| 326 |
+
"step": 775
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 10.0,
|
| 330 |
+
"grad_norm": 0.3718573749065399,
|
| 331 |
+
"learning_rate": 1.1125000000000001e-05,
|
| 332 |
+
"loss": 0.1696,
|
| 333 |
+
"step": 800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 10.0,
|
| 337 |
+
"eval_accuracy": 0.897887323943662,
|
| 338 |
+
"eval_loss": 0.6130573153495789,
|
| 339 |
+
"eval_macro_f1": 0.882421802504033,
|
| 340 |
+
"eval_runtime": 0.2612,
|
| 341 |
+
"eval_samples_per_second": 1087.48,
|
| 342 |
+
"eval_steps_per_second": 68.925,
|
| 343 |
+
"eval_weighted_f1": 0.8974368718628961,
|
| 344 |
+
"step": 800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 10.3125,
|
| 348 |
+
"grad_norm": 133.16798400878906,
|
| 349 |
+
"learning_rate": 1.0777777777777778e-05,
|
| 350 |
+
"loss": 0.3378,
|
| 351 |
+
"step": 825
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 10.625,
|
| 355 |
+
"grad_norm": 3.597330331802368,
|
| 356 |
+
"learning_rate": 1.0430555555555557e-05,
|
| 357 |
+
"loss": 0.3234,
|
| 358 |
+
"step": 850
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 10.9375,
|
| 362 |
+
"grad_norm": 115.65129852294922,
|
| 363 |
+
"learning_rate": 1.0083333333333334e-05,
|
| 364 |
+
"loss": 0.2238,
|
| 365 |
+
"step": 875
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 11.0,
|
| 369 |
+
"eval_accuracy": 0.8802816901408451,
|
| 370 |
+
"eval_loss": 0.6647355556488037,
|
| 371 |
+
"eval_macro_f1": 0.8563095238095239,
|
| 372 |
+
"eval_runtime": 0.2674,
|
| 373 |
+
"eval_samples_per_second": 1062.104,
|
| 374 |
+
"eval_steps_per_second": 67.316,
|
| 375 |
+
"eval_weighted_f1": 0.8769751844399732,
|
| 376 |
+
"step": 880
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 11.25,
|
| 380 |
+
"grad_norm": 0.06744170933961868,
|
| 381 |
+
"learning_rate": 9.736111111111112e-06,
|
| 382 |
+
"loss": 0.0535,
|
| 383 |
+
"step": 900
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 11.5625,
|
| 387 |
+
"grad_norm": 0.04192639887332916,
|
| 388 |
+
"learning_rate": 9.38888888888889e-06,
|
| 389 |
+
"loss": 0.2063,
|
| 390 |
+
"step": 925
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 11.875,
|
| 394 |
+
"grad_norm": 226.21644592285156,
|
| 395 |
+
"learning_rate": 9.041666666666667e-06,
|
| 396 |
+
"loss": 0.2842,
|
| 397 |
+
"step": 950
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 12.0,
|
| 401 |
+
"eval_accuracy": 0.8908450704225352,
|
| 402 |
+
"eval_loss": 0.685333788394928,
|
| 403 |
+
"eval_macro_f1": 0.8757462423258768,
|
| 404 |
+
"eval_runtime": 0.2615,
|
| 405 |
+
"eval_samples_per_second": 1085.841,
|
| 406 |
+
"eval_steps_per_second": 68.821,
|
| 407 |
+
"eval_weighted_f1": 0.890997583837653,
|
| 408 |
+
"step": 960
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 12.1875,
|
| 412 |
+
"grad_norm": 0.12042868137359619,
|
| 413 |
+
"learning_rate": 8.694444444444445e-06,
|
| 414 |
+
"loss": 0.1968,
|
| 415 |
+
"step": 975
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 12.5,
|
| 419 |
+
"grad_norm": 198.77696228027344,
|
| 420 |
+
"learning_rate": 8.347222222222223e-06,
|
| 421 |
+
"loss": 0.1945,
|
| 422 |
+
"step": 1000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 12.8125,
|
| 426 |
+
"grad_norm": 2.0745413303375244,
|
| 427 |
+
"learning_rate": 8.000000000000001e-06,
|
| 428 |
+
"loss": 0.2213,
|
| 429 |
+
"step": 1025
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 13.0,
|
| 433 |
+
"eval_accuracy": 0.8591549295774648,
|
| 434 |
+
"eval_loss": 0.8092403411865234,
|
| 435 |
+
"eval_macro_f1": 0.8246372337141092,
|
| 436 |
+
"eval_runtime": 0.2722,
|
| 437 |
+
"eval_samples_per_second": 1043.482,
|
| 438 |
+
"eval_steps_per_second": 66.136,
|
| 439 |
+
"eval_weighted_f1": 0.8520322304310581,
|
| 440 |
+
"step": 1040
|
| 441 |
+
}
|
| 442 |
+
],
|
| 443 |
+
"logging_steps": 25,
|
| 444 |
+
"max_steps": 1600,
|
| 445 |
+
"num_input_tokens_seen": 0,
|
| 446 |
+
"num_train_epochs": 20,
|
| 447 |
+
"save_steps": 500,
|
| 448 |
+
"stateful_callbacks": {
|
| 449 |
+
"EarlyStoppingCallback": {
|
| 450 |
+
"args": {
|
| 451 |
+
"early_stopping_patience": 3,
|
| 452 |
+
"early_stopping_threshold": 0.0
|
| 453 |
+
},
|
| 454 |
+
"attributes": {
|
| 455 |
+
"early_stopping_patience_counter": 3
|
| 456 |
+
}
|
| 457 |
+
},
|
| 458 |
+
"TrainerControl": {
|
| 459 |
+
"args": {
|
| 460 |
+
"should_epoch_stop": false,
|
| 461 |
+
"should_evaluate": false,
|
| 462 |
+
"should_log": false,
|
| 463 |
+
"should_save": true,
|
| 464 |
+
"should_training_stop": true
|
| 465 |
+
},
|
| 466 |
+
"attributes": {}
|
| 467 |
+
}
|
| 468 |
+
},
|
| 469 |
+
"total_flos": 0.0,
|
| 470 |
+
"train_batch_size": 8,
|
| 471 |
+
"trial_name": null,
|
| 472 |
+
"trial_params": null
|
| 473 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800/trainer_state.json
ADDED
|
@@ -0,0 +1,377 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 800,
|
| 3 |
+
"best_metric": 0.882421802504033,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-04-s44-20260827T174832Z-fe731097/trainer/checkpoint-800",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 800,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.3125,
|
| 14 |
+
"grad_norm": 17.846830368041992,
|
| 15 |
+
"learning_rate": 3e-06,
|
| 16 |
+
"loss": 2.5728,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.625,
|
| 21 |
+
"grad_norm": 11.913187980651855,
|
| 22 |
+
"learning_rate": 6.125000000000001e-06,
|
| 23 |
+
"loss": 2.5396,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.9375,
|
| 28 |
+
"grad_norm": 9.724447250366211,
|
| 29 |
+
"learning_rate": 9.250000000000001e-06,
|
| 30 |
+
"loss": 2.4403,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.676056338028169,
|
| 36 |
+
"eval_loss": 0.5894998908042908,
|
| 37 |
+
"eval_macro_f1": 0.40336134453781514,
|
| 38 |
+
"eval_runtime": 0.2623,
|
| 39 |
+
"eval_samples_per_second": 1082.721,
|
| 40 |
+
"eval_steps_per_second": 68.623,
|
| 41 |
+
"eval_weighted_f1": 0.5453899869807078,
|
| 42 |
+
"step": 80
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.25,
|
| 46 |
+
"grad_norm": 27.37001609802246,
|
| 47 |
+
"learning_rate": 1.2375000000000001e-05,
|
| 48 |
+
"loss": 2.2358,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.5625,
|
| 53 |
+
"grad_norm": 24.617549896240234,
|
| 54 |
+
"learning_rate": 1.55e-05,
|
| 55 |
+
"loss": 2.0279,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.875,
|
| 60 |
+
"grad_norm": 24.2894344329834,
|
| 61 |
+
"learning_rate": 1.8625000000000002e-05,
|
| 62 |
+
"loss": 1.9649,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8415492957746479,
|
| 68 |
+
"eval_loss": 0.38190901279449463,
|
| 69 |
+
"eval_macro_f1": 0.8129801712153362,
|
| 70 |
+
"eval_runtime": 0.2571,
|
| 71 |
+
"eval_samples_per_second": 1104.709,
|
| 72 |
+
"eval_steps_per_second": 70.017,
|
| 73 |
+
"eval_weighted_f1": 0.8387181212687701,
|
| 74 |
+
"step": 160
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.1875,
|
| 78 |
+
"grad_norm": 86.57966613769531,
|
| 79 |
+
"learning_rate": 1.9805555555555557e-05,
|
| 80 |
+
"loss": 1.6063,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.5,
|
| 85 |
+
"grad_norm": 45.83436965942383,
|
| 86 |
+
"learning_rate": 1.9458333333333333e-05,
|
| 87 |
+
"loss": 1.6571,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.8125,
|
| 92 |
+
"grad_norm": 49.14204406738281,
|
| 93 |
+
"learning_rate": 1.9111111111111113e-05,
|
| 94 |
+
"loss": 1.5296,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8732394366197183,
|
| 100 |
+
"eval_loss": 0.3036462366580963,
|
| 101 |
+
"eval_macro_f1": 0.8536082474226805,
|
| 102 |
+
"eval_runtime": 0.2673,
|
| 103 |
+
"eval_samples_per_second": 1062.583,
|
| 104 |
+
"eval_steps_per_second": 67.347,
|
| 105 |
+
"eval_weighted_f1": 0.8724843908813708,
|
| 106 |
+
"step": 240
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.125,
|
| 110 |
+
"grad_norm": 41.95870590209961,
|
| 111 |
+
"learning_rate": 1.876388888888889e-05,
|
| 112 |
+
"loss": 1.3436,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.4375,
|
| 117 |
+
"grad_norm": 26.975175857543945,
|
| 118 |
+
"learning_rate": 1.8416666666666666e-05,
|
| 119 |
+
"loss": 1.0426,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.75,
|
| 124 |
+
"grad_norm": 36.39384841918945,
|
| 125 |
+
"learning_rate": 1.8069444444444446e-05,
|
| 126 |
+
"loss": 1.2598,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8661971830985915,
|
| 132 |
+
"eval_loss": 0.32074618339538574,
|
| 133 |
+
"eval_macro_f1": 0.8359180198260658,
|
| 134 |
+
"eval_runtime": 0.262,
|
| 135 |
+
"eval_samples_per_second": 1084.045,
|
| 136 |
+
"eval_steps_per_second": 68.707,
|
| 137 |
+
"eval_weighted_f1": 0.8607370061150214,
|
| 138 |
+
"step": 320
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.0625,
|
| 142 |
+
"grad_norm": 34.646976470947266,
|
| 143 |
+
"learning_rate": 1.7722222222222222e-05,
|
| 144 |
+
"loss": 1.139,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.375,
|
| 149 |
+
"grad_norm": 53.85174560546875,
|
| 150 |
+
"learning_rate": 1.7375000000000002e-05,
|
| 151 |
+
"loss": 0.6466,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.6875,
|
| 156 |
+
"grad_norm": 39.1696891784668,
|
| 157 |
+
"learning_rate": 1.702777777777778e-05,
|
| 158 |
+
"loss": 0.9732,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"grad_norm": 67.7001724243164,
|
| 164 |
+
"learning_rate": 1.668055555555556e-05,
|
| 165 |
+
"loss": 1.0013,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 5.0,
|
| 170 |
+
"eval_accuracy": 0.8802816901408451,
|
| 171 |
+
"eval_loss": 0.34587180614471436,
|
| 172 |
+
"eval_macro_f1": 0.857286432160804,
|
| 173 |
+
"eval_runtime": 0.2552,
|
| 174 |
+
"eval_samples_per_second": 1112.745,
|
| 175 |
+
"eval_steps_per_second": 70.526,
|
| 176 |
+
"eval_weighted_f1": 0.8774577110906647,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.3125,
|
| 181 |
+
"grad_norm": 99.6565170288086,
|
| 182 |
+
"learning_rate": 1.6333333333333335e-05,
|
| 183 |
+
"loss": 0.7472,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 5.625,
|
| 188 |
+
"grad_norm": 11.587423324584961,
|
| 189 |
+
"learning_rate": 1.5986111111111115e-05,
|
| 190 |
+
"loss": 0.7773,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.9375,
|
| 195 |
+
"grad_norm": 27.089900970458984,
|
| 196 |
+
"learning_rate": 1.563888888888889e-05,
|
| 197 |
+
"loss": 0.4622,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 6.0,
|
| 202 |
+
"eval_accuracy": 0.8415492957746479,
|
| 203 |
+
"eval_loss": 0.6065791845321655,
|
| 204 |
+
"eval_macro_f1": 0.7897161661867544,
|
| 205 |
+
"eval_runtime": 0.3906,
|
| 206 |
+
"eval_samples_per_second": 727.002,
|
| 207 |
+
"eval_steps_per_second": 46.078,
|
| 208 |
+
"eval_weighted_f1": 0.8264772510008633,
|
| 209 |
+
"step": 480
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.25,
|
| 213 |
+
"grad_norm": 46.45066833496094,
|
| 214 |
+
"learning_rate": 1.5291666666666668e-05,
|
| 215 |
+
"loss": 0.5752,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 6.5625,
|
| 220 |
+
"grad_norm": 39.581424713134766,
|
| 221 |
+
"learning_rate": 1.4944444444444446e-05,
|
| 222 |
+
"loss": 0.6317,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.875,
|
| 227 |
+
"grad_norm": 22.54334831237793,
|
| 228 |
+
"learning_rate": 1.4597222222222223e-05,
|
| 229 |
+
"loss": 0.3716,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 7.0,
|
| 234 |
+
"eval_accuracy": 0.8556338028169014,
|
| 235 |
+
"eval_loss": 0.5141348242759705,
|
| 236 |
+
"eval_macro_f1": 0.8391157167530225,
|
| 237 |
+
"eval_runtime": 0.2782,
|
| 238 |
+
"eval_samples_per_second": 1020.909,
|
| 239 |
+
"eval_steps_per_second": 64.706,
|
| 240 |
+
"eval_weighted_f1": 0.857267459680362,
|
| 241 |
+
"step": 560
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.1875,
|
| 245 |
+
"grad_norm": 80.10563659667969,
|
| 246 |
+
"learning_rate": 1.425e-05,
|
| 247 |
+
"loss": 0.4154,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 7.5,
|
| 252 |
+
"grad_norm": 34.158931732177734,
|
| 253 |
+
"learning_rate": 1.3902777777777779e-05,
|
| 254 |
+
"loss": 0.4169,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 7.8125,
|
| 259 |
+
"grad_norm": 4.919360637664795,
|
| 260 |
+
"learning_rate": 1.3555555555555557e-05,
|
| 261 |
+
"loss": 0.3483,
|
| 262 |
+
"step": 625
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 8.0,
|
| 266 |
+
"eval_accuracy": 0.8802816901408451,
|
| 267 |
+
"eval_loss": 0.5978385210037231,
|
| 268 |
+
"eval_macro_f1": 0.8582334977683814,
|
| 269 |
+
"eval_runtime": 0.2702,
|
| 270 |
+
"eval_samples_per_second": 1050.961,
|
| 271 |
+
"eval_steps_per_second": 66.61,
|
| 272 |
+
"eval_weighted_f1": 0.8779193838152238,
|
| 273 |
+
"step": 640
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.125,
|
| 277 |
+
"grad_norm": 7.283858299255371,
|
| 278 |
+
"learning_rate": 1.3208333333333334e-05,
|
| 279 |
+
"loss": 0.3996,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 8.4375,
|
| 284 |
+
"grad_norm": 7.524752140045166,
|
| 285 |
+
"learning_rate": 1.2861111111111114e-05,
|
| 286 |
+
"loss": 0.3507,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 8.75,
|
| 291 |
+
"grad_norm": 83.3502197265625,
|
| 292 |
+
"learning_rate": 1.251388888888889e-05,
|
| 293 |
+
"loss": 0.2851,
|
| 294 |
+
"step": 700
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 9.0,
|
| 298 |
+
"eval_accuracy": 0.8767605633802817,
|
| 299 |
+
"eval_loss": 0.6831046342849731,
|
| 300 |
+
"eval_macro_f1": 0.855470737913486,
|
| 301 |
+
"eval_runtime": 0.2725,
|
| 302 |
+
"eval_samples_per_second": 1042.115,
|
| 303 |
+
"eval_steps_per_second": 66.05,
|
| 304 |
+
"eval_weighted_f1": 0.8750026878830235,
|
| 305 |
+
"step": 720
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 9.0625,
|
| 309 |
+
"grad_norm": 0.4345070421695709,
|
| 310 |
+
"learning_rate": 1.2166666666666667e-05,
|
| 311 |
+
"loss": 0.3616,
|
| 312 |
+
"step": 725
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 9.375,
|
| 316 |
+
"grad_norm": 0.09883200377225876,
|
| 317 |
+
"learning_rate": 1.1819444444444446e-05,
|
| 318 |
+
"loss": 0.2566,
|
| 319 |
+
"step": 750
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 9.6875,
|
| 323 |
+
"grad_norm": 56.23919677734375,
|
| 324 |
+
"learning_rate": 1.1472222222222223e-05,
|
| 325 |
+
"loss": 0.3902,
|
| 326 |
+
"step": 775
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 10.0,
|
| 330 |
+
"grad_norm": 0.3718573749065399,
|
| 331 |
+
"learning_rate": 1.1125000000000001e-05,
|
| 332 |
+
"loss": 0.1696,
|
| 333 |
+
"step": 800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 10.0,
|
| 337 |
+
"eval_accuracy": 0.897887323943662,
|
| 338 |
+
"eval_loss": 0.6130573153495789,
|
| 339 |
+
"eval_macro_f1": 0.882421802504033,
|
| 340 |
+
"eval_runtime": 0.2612,
|
| 341 |
+
"eval_samples_per_second": 1087.48,
|
| 342 |
+
"eval_steps_per_second": 68.925,
|
| 343 |
+
"eval_weighted_f1": 0.8974368718628961,
|
| 344 |
+
"step": 800
|
| 345 |
+
}
|
| 346 |
+
],
|
| 347 |
+
"logging_steps": 25,
|
| 348 |
+
"max_steps": 1600,
|
| 349 |
+
"num_input_tokens_seen": 0,
|
| 350 |
+
"num_train_epochs": 20,
|
| 351 |
+
"save_steps": 500,
|
| 352 |
+
"stateful_callbacks": {
|
| 353 |
+
"EarlyStoppingCallback": {
|
| 354 |
+
"args": {
|
| 355 |
+
"early_stopping_patience": 3,
|
| 356 |
+
"early_stopping_threshold": 0.0
|
| 357 |
+
},
|
| 358 |
+
"attributes": {
|
| 359 |
+
"early_stopping_patience_counter": 0
|
| 360 |
+
}
|
| 361 |
+
},
|
| 362 |
+
"TrainerControl": {
|
| 363 |
+
"args": {
|
| 364 |
+
"should_epoch_stop": false,
|
| 365 |
+
"should_evaluate": false,
|
| 366 |
+
"should_log": false,
|
| 367 |
+
"should_save": true,
|
| 368 |
+
"should_training_stop": false
|
| 369 |
+
},
|
| 370 |
+
"attributes": {}
|
| 371 |
+
}
|
| 372 |
+
},
|
| 373 |
+
"total_flos": 0.0,
|
| 374 |
+
"train_batch_size": 8,
|
| 375 |
+
"trial_name": null,
|
| 376 |
+
"trial_params": null
|
| 377 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s43-20260827T173648Z-32d744c6/trainer/checkpoint-810/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-05-s43-20260827T173648Z-32d744c6/trainer/checkpoint-810/trainer_state.json
ADDED
|
@@ -0,0 +1,366 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 540,
|
| 3 |
+
"best_metric": 0.8536785814804391,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-05-s43-20260827T173648Z-32d744c6/trainer/checkpoint-540",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 810,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2801120448179272,
|
| 14 |
+
"grad_norm": 35.33953094482422,
|
| 15 |
+
"learning_rate": 2.666666666666667e-06,
|
| 16 |
+
"loss": 2.8117,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5602240896358543,
|
| 21 |
+
"grad_norm": 22.959726333618164,
|
| 22 |
+
"learning_rate": 5.444444444444445e-06,
|
| 23 |
+
"loss": 2.4573,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.8403361344537815,
|
| 28 |
+
"grad_norm": 40.109745025634766,
|
| 29 |
+
"learning_rate": 8.222222222222222e-06,
|
| 30 |
+
"loss": 2.5642,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6466876971608833,
|
| 36 |
+
"eval_loss": 0.6363799571990967,
|
| 37 |
+
"eval_macro_f1": 0.39272030651340994,
|
| 38 |
+
"eval_runtime": 0.2727,
|
| 39 |
+
"eval_samples_per_second": 1162.551,
|
| 40 |
+
"eval_steps_per_second": 73.347,
|
| 41 |
+
"eval_weighted_f1": 0.5079347812949466,
|
| 42 |
+
"step": 90
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.112044817927171,
|
| 46 |
+
"grad_norm": 23.535064697265625,
|
| 47 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 48 |
+
"loss": 2.3463,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.392156862745098,
|
| 53 |
+
"grad_norm": 27.682165145874023,
|
| 54 |
+
"learning_rate": 1.377777777777778e-05,
|
| 55 |
+
"loss": 2.1641,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.6722689075630253,
|
| 60 |
+
"grad_norm": 28.0634765625,
|
| 61 |
+
"learning_rate": 1.6555555555555556e-05,
|
| 62 |
+
"loss": 1.9028,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.9523809523809523,
|
| 67 |
+
"grad_norm": 30.133153915405273,
|
| 68 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 69 |
+
"loss": 1.7868,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8264984227129337,
|
| 75 |
+
"eval_loss": 0.44510340690612793,
|
| 76 |
+
"eval_macro_f1": 0.7901041353157165,
|
| 77 |
+
"eval_runtime": 0.2852,
|
| 78 |
+
"eval_samples_per_second": 1111.364,
|
| 79 |
+
"eval_steps_per_second": 70.118,
|
| 80 |
+
"eval_weighted_f1": 0.815745565072847,
|
| 81 |
+
"step": 180
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.224089635854342,
|
| 85 |
+
"grad_norm": 47.92458724975586,
|
| 86 |
+
"learning_rate": 1.9765432098765434e-05,
|
| 87 |
+
"loss": 1.4243,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.504201680672269,
|
| 92 |
+
"grad_norm": 38.56916046142578,
|
| 93 |
+
"learning_rate": 1.9456790123456793e-05,
|
| 94 |
+
"loss": 1.4229,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.784313725490196,
|
| 99 |
+
"grad_norm": 43.35906982421875,
|
| 100 |
+
"learning_rate": 1.9148148148148148e-05,
|
| 101 |
+
"loss": 1.4901,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_accuracy": 0.8264984227129337,
|
| 107 |
+
"eval_loss": 0.39202427864074707,
|
| 108 |
+
"eval_macro_f1": 0.8037151702786378,
|
| 109 |
+
"eval_runtime": 0.2681,
|
| 110 |
+
"eval_samples_per_second": 1182.34,
|
| 111 |
+
"eval_steps_per_second": 74.596,
|
| 112 |
+
"eval_weighted_f1": 0.8233340820970593,
|
| 113 |
+
"step": 270
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.0560224089635852,
|
| 117 |
+
"grad_norm": 31.804914474487305,
|
| 118 |
+
"learning_rate": 1.8839506172839507e-05,
|
| 119 |
+
"loss": 1.2323,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.3361344537815127,
|
| 124 |
+
"grad_norm": 36.353271484375,
|
| 125 |
+
"learning_rate": 1.8530864197530866e-05,
|
| 126 |
+
"loss": 1.101,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.61624649859944,
|
| 131 |
+
"grad_norm": 21.716171264648438,
|
| 132 |
+
"learning_rate": 1.8222222222222224e-05,
|
| 133 |
+
"loss": 1.0831,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.896358543417367,
|
| 138 |
+
"grad_norm": 46.629600524902344,
|
| 139 |
+
"learning_rate": 1.7913580246913583e-05,
|
| 140 |
+
"loss": 1.1386,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.8485804416403786,
|
| 146 |
+
"eval_loss": 0.37881559133529663,
|
| 147 |
+
"eval_macro_f1": 0.82911051212938,
|
| 148 |
+
"eval_runtime": 0.2947,
|
| 149 |
+
"eval_samples_per_second": 1075.633,
|
| 150 |
+
"eval_steps_per_second": 67.863,
|
| 151 |
+
"eval_weighted_f1": 0.8460329742277246,
|
| 152 |
+
"step": 360
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.168067226890757,
|
| 156 |
+
"grad_norm": 47.786495208740234,
|
| 157 |
+
"learning_rate": 1.760493827160494e-05,
|
| 158 |
+
"loss": 0.7662,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.448179271708684,
|
| 163 |
+
"grad_norm": 56.675933837890625,
|
| 164 |
+
"learning_rate": 1.7296296296296297e-05,
|
| 165 |
+
"loss": 0.819,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.728291316526611,
|
| 170 |
+
"grad_norm": 30.737823486328125,
|
| 171 |
+
"learning_rate": 1.6987654320987656e-05,
|
| 172 |
+
"loss": 0.7683,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 5.0,
|
| 177 |
+
"grad_norm": 1.362674355506897,
|
| 178 |
+
"learning_rate": 1.6679012345679015e-05,
|
| 179 |
+
"loss": 0.6083,
|
| 180 |
+
"step": 450
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.0,
|
| 184 |
+
"eval_accuracy": 0.8138801261829653,
|
| 185 |
+
"eval_loss": 0.6949102878570557,
|
| 186 |
+
"eval_macro_f1": 0.7703603658910921,
|
| 187 |
+
"eval_runtime": 0.3215,
|
| 188 |
+
"eval_samples_per_second": 985.859,
|
| 189 |
+
"eval_steps_per_second": 62.199,
|
| 190 |
+
"eval_weighted_f1": 0.7996889000008327,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.280112044817927,
|
| 195 |
+
"grad_norm": 96.77898406982422,
|
| 196 |
+
"learning_rate": 1.6370370370370374e-05,
|
| 197 |
+
"loss": 0.555,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.560224089635854,
|
| 202 |
+
"grad_norm": 25.191293716430664,
|
| 203 |
+
"learning_rate": 1.606172839506173e-05,
|
| 204 |
+
"loss": 0.4355,
|
| 205 |
+
"step": 500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 5.840336134453781,
|
| 209 |
+
"grad_norm": 42.99637985229492,
|
| 210 |
+
"learning_rate": 1.5753086419753088e-05,
|
| 211 |
+
"loss": 0.6684,
|
| 212 |
+
"step": 525
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 6.0,
|
| 216 |
+
"eval_accuracy": 0.8706624605678234,
|
| 217 |
+
"eval_loss": 0.4825526177883148,
|
| 218 |
+
"eval_macro_f1": 0.8536785814804391,
|
| 219 |
+
"eval_runtime": 0.3002,
|
| 220 |
+
"eval_samples_per_second": 1056.03,
|
| 221 |
+
"eval_steps_per_second": 66.626,
|
| 222 |
+
"eval_weighted_f1": 0.8683035884723532,
|
| 223 |
+
"step": 540
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.1120448179271705,
|
| 227 |
+
"grad_norm": 53.533260345458984,
|
| 228 |
+
"learning_rate": 1.5444444444444446e-05,
|
| 229 |
+
"loss": 0.4718,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 6.392156862745098,
|
| 234 |
+
"grad_norm": 96.22107696533203,
|
| 235 |
+
"learning_rate": 1.5135802469135803e-05,
|
| 236 |
+
"loss": 0.5424,
|
| 237 |
+
"step": 575
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 6.6722689075630255,
|
| 241 |
+
"grad_norm": 34.11431121826172,
|
| 242 |
+
"learning_rate": 1.4827160493827162e-05,
|
| 243 |
+
"loss": 0.4675,
|
| 244 |
+
"step": 600
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 6.9523809523809526,
|
| 248 |
+
"grad_norm": 19.950820922851562,
|
| 249 |
+
"learning_rate": 1.4518518518518521e-05,
|
| 250 |
+
"loss": 0.4186,
|
| 251 |
+
"step": 625
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 7.0,
|
| 255 |
+
"eval_accuracy": 0.8517350157728707,
|
| 256 |
+
"eval_loss": 0.6454525589942932,
|
| 257 |
+
"eval_macro_f1": 0.8249503601095015,
|
| 258 |
+
"eval_runtime": 0.2683,
|
| 259 |
+
"eval_samples_per_second": 1181.37,
|
| 260 |
+
"eval_steps_per_second": 74.534,
|
| 261 |
+
"eval_weighted_f1": 0.8450388518570283,
|
| 262 |
+
"step": 630
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 7.224089635854342,
|
| 266 |
+
"grad_norm": 45.37995147705078,
|
| 267 |
+
"learning_rate": 1.4209876543209878e-05,
|
| 268 |
+
"loss": 0.2373,
|
| 269 |
+
"step": 650
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 7.504201680672269,
|
| 273 |
+
"grad_norm": 59.21278381347656,
|
| 274 |
+
"learning_rate": 1.3901234567901237e-05,
|
| 275 |
+
"loss": 0.6121,
|
| 276 |
+
"step": 675
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 7.784313725490196,
|
| 280 |
+
"grad_norm": 0.4692898690700531,
|
| 281 |
+
"learning_rate": 1.3592592592592592e-05,
|
| 282 |
+
"loss": 0.2202,
|
| 283 |
+
"step": 700
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 8.0,
|
| 287 |
+
"eval_accuracy": 0.8548895899053628,
|
| 288 |
+
"eval_loss": 0.6984889507293701,
|
| 289 |
+
"eval_macro_f1": 0.8354324665944384,
|
| 290 |
+
"eval_runtime": 0.2702,
|
| 291 |
+
"eval_samples_per_second": 1173.111,
|
| 292 |
+
"eval_steps_per_second": 74.013,
|
| 293 |
+
"eval_weighted_f1": 0.8520334984101811,
|
| 294 |
+
"step": 720
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 8.056022408963585,
|
| 298 |
+
"grad_norm": 3.2777211666107178,
|
| 299 |
+
"learning_rate": 1.3283950617283951e-05,
|
| 300 |
+
"loss": 0.2828,
|
| 301 |
+
"step": 725
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 8.336134453781513,
|
| 305 |
+
"grad_norm": 0.36308854818344116,
|
| 306 |
+
"learning_rate": 1.297530864197531e-05,
|
| 307 |
+
"loss": 0.4868,
|
| 308 |
+
"step": 750
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 8.61624649859944,
|
| 312 |
+
"grad_norm": 1.882088541984558,
|
| 313 |
+
"learning_rate": 1.2666666666666667e-05,
|
| 314 |
+
"loss": 0.4352,
|
| 315 |
+
"step": 775
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 8.896358543417367,
|
| 319 |
+
"grad_norm": 59.212852478027344,
|
| 320 |
+
"learning_rate": 1.2358024691358026e-05,
|
| 321 |
+
"loss": 0.4002,
|
| 322 |
+
"step": 800
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 9.0,
|
| 326 |
+
"eval_accuracy": 0.8517350157728707,
|
| 327 |
+
"eval_loss": 0.8467510938644409,
|
| 328 |
+
"eval_macro_f1": 0.8269349161913833,
|
| 329 |
+
"eval_runtime": 0.2743,
|
| 330 |
+
"eval_samples_per_second": 1155.646,
|
| 331 |
+
"eval_steps_per_second": 72.911,
|
| 332 |
+
"eval_weighted_f1": 0.8461549933670361,
|
| 333 |
+
"step": 810
|
| 334 |
+
}
|
| 335 |
+
],
|
| 336 |
+
"logging_steps": 25,
|
| 337 |
+
"max_steps": 1800,
|
| 338 |
+
"num_input_tokens_seen": 0,
|
| 339 |
+
"num_train_epochs": 20,
|
| 340 |
+
"save_steps": 500,
|
| 341 |
+
"stateful_callbacks": {
|
| 342 |
+
"EarlyStoppingCallback": {
|
| 343 |
+
"args": {
|
| 344 |
+
"early_stopping_patience": 3,
|
| 345 |
+
"early_stopping_threshold": 0.0
|
| 346 |
+
},
|
| 347 |
+
"attributes": {
|
| 348 |
+
"early_stopping_patience_counter": 3
|
| 349 |
+
}
|
| 350 |
+
},
|
| 351 |
+
"TrainerControl": {
|
| 352 |
+
"args": {
|
| 353 |
+
"should_epoch_stop": false,
|
| 354 |
+
"should_evaluate": false,
|
| 355 |
+
"should_log": false,
|
| 356 |
+
"should_save": true,
|
| 357 |
+
"should_training_stop": true
|
| 358 |
+
},
|
| 359 |
+
"attributes": {}
|
| 360 |
+
}
|
| 361 |
+
},
|
| 362 |
+
"total_flos": 0.0,
|
| 363 |
+
"train_batch_size": 8,
|
| 364 |
+
"trial_name": null,
|
| 365 |
+
"trial_params": null
|
| 366 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450/trainer_state.json
ADDED
|
@@ -0,0 +1,224 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 450,
|
| 3 |
+
"best_metric": 0.8334384195039932,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 450,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2801120448179272,
|
| 14 |
+
"grad_norm": 28.82010269165039,
|
| 15 |
+
"learning_rate": 2.666666666666667e-06,
|
| 16 |
+
"loss": 2.5697,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5602240896358543,
|
| 21 |
+
"grad_norm": 7.371585845947266,
|
| 22 |
+
"learning_rate": 5.444444444444445e-06,
|
| 23 |
+
"loss": 2.5197,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.8403361344537815,
|
| 28 |
+
"grad_norm": 15.176864624023438,
|
| 29 |
+
"learning_rate": 8.222222222222222e-06,
|
| 30 |
+
"loss": 2.4669,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6656151419558359,
|
| 36 |
+
"eval_loss": 0.5750473737716675,
|
| 37 |
+
"eval_macro_f1": 0.45522049286640726,
|
| 38 |
+
"eval_runtime": 0.2786,
|
| 39 |
+
"eval_samples_per_second": 1137.717,
|
| 40 |
+
"eval_steps_per_second": 71.78,
|
| 41 |
+
"eval_weighted_f1": 0.554543855126899,
|
| 42 |
+
"step": 90
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.112044817927171,
|
| 46 |
+
"grad_norm": 21.1071720123291,
|
| 47 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 48 |
+
"loss": 2.2098,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.392156862745098,
|
| 53 |
+
"grad_norm": 37.07575988769531,
|
| 54 |
+
"learning_rate": 1.377777777777778e-05,
|
| 55 |
+
"loss": 1.9603,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.6722689075630253,
|
| 60 |
+
"grad_norm": 27.69922637939453,
|
| 61 |
+
"learning_rate": 1.6555555555555556e-05,
|
| 62 |
+
"loss": 2.0119,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.9523809523809523,
|
| 67 |
+
"grad_norm": 35.82430648803711,
|
| 68 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 69 |
+
"loss": 1.7122,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8012618296529969,
|
| 75 |
+
"eval_loss": 0.4539729058742523,
|
| 76 |
+
"eval_macro_f1": 0.7639724392232872,
|
| 77 |
+
"eval_runtime": 0.3245,
|
| 78 |
+
"eval_samples_per_second": 976.963,
|
| 79 |
+
"eval_steps_per_second": 61.638,
|
| 80 |
+
"eval_weighted_f1": 0.79149556073093,
|
| 81 |
+
"step": 180
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.224089635854342,
|
| 85 |
+
"grad_norm": 42.21854019165039,
|
| 86 |
+
"learning_rate": 1.9765432098765434e-05,
|
| 87 |
+
"loss": 1.3959,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.504201680672269,
|
| 92 |
+
"grad_norm": 27.365318298339844,
|
| 93 |
+
"learning_rate": 1.9456790123456793e-05,
|
| 94 |
+
"loss": 1.6084,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.784313725490196,
|
| 99 |
+
"grad_norm": 47.25198745727539,
|
| 100 |
+
"learning_rate": 1.9148148148148148e-05,
|
| 101 |
+
"loss": 1.4331,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_accuracy": 0.8422712933753943,
|
| 107 |
+
"eval_loss": 0.3598579168319702,
|
| 108 |
+
"eval_macro_f1": 0.8236537605696483,
|
| 109 |
+
"eval_runtime": 0.2716,
|
| 110 |
+
"eval_samples_per_second": 1167.18,
|
| 111 |
+
"eval_steps_per_second": 73.639,
|
| 112 |
+
"eval_weighted_f1": 0.8404637659185258,
|
| 113 |
+
"step": 270
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.0560224089635852,
|
| 117 |
+
"grad_norm": 38.71910858154297,
|
| 118 |
+
"learning_rate": 1.8839506172839507e-05,
|
| 119 |
+
"loss": 1.3697,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.3361344537815127,
|
| 124 |
+
"grad_norm": 30.123592376708984,
|
| 125 |
+
"learning_rate": 1.8530864197530866e-05,
|
| 126 |
+
"loss": 1.1044,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.61624649859944,
|
| 131 |
+
"grad_norm": 50.46794891357422,
|
| 132 |
+
"learning_rate": 1.8222222222222224e-05,
|
| 133 |
+
"loss": 0.9246,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.896358543417367,
|
| 138 |
+
"grad_norm": 33.4803466796875,
|
| 139 |
+
"learning_rate": 1.7913580246913583e-05,
|
| 140 |
+
"loss": 1.1125,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.8201892744479495,
|
| 146 |
+
"eval_loss": 0.5016289949417114,
|
| 147 |
+
"eval_macro_f1": 0.7877057558774805,
|
| 148 |
+
"eval_runtime": 0.2859,
|
| 149 |
+
"eval_samples_per_second": 1108.765,
|
| 150 |
+
"eval_steps_per_second": 69.954,
|
| 151 |
+
"eval_weighted_f1": 0.8120683948053322,
|
| 152 |
+
"step": 360
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.168067226890757,
|
| 156 |
+
"grad_norm": 29.374866485595703,
|
| 157 |
+
"learning_rate": 1.760493827160494e-05,
|
| 158 |
+
"loss": 0.8453,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.448179271708684,
|
| 163 |
+
"grad_norm": 44.94426727294922,
|
| 164 |
+
"learning_rate": 1.7296296296296297e-05,
|
| 165 |
+
"loss": 0.8138,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.728291316526611,
|
| 170 |
+
"grad_norm": 38.09920883178711,
|
| 171 |
+
"learning_rate": 1.6987654320987656e-05,
|
| 172 |
+
"loss": 0.7823,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 5.0,
|
| 177 |
+
"grad_norm": 4.866731643676758,
|
| 178 |
+
"learning_rate": 1.6679012345679015e-05,
|
| 179 |
+
"loss": 0.767,
|
| 180 |
+
"step": 450
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.0,
|
| 184 |
+
"eval_accuracy": 0.8422712933753943,
|
| 185 |
+
"eval_loss": 0.3990023732185364,
|
| 186 |
+
"eval_macro_f1": 0.8334384195039932,
|
| 187 |
+
"eval_runtime": 0.2885,
|
| 188 |
+
"eval_samples_per_second": 1098.655,
|
| 189 |
+
"eval_steps_per_second": 69.316,
|
| 190 |
+
"eval_weighted_f1": 0.8446912588196138,
|
| 191 |
+
"step": 450
|
| 192 |
+
}
|
| 193 |
+
],
|
| 194 |
+
"logging_steps": 25,
|
| 195 |
+
"max_steps": 1800,
|
| 196 |
+
"num_input_tokens_seen": 0,
|
| 197 |
+
"num_train_epochs": 20,
|
| 198 |
+
"save_steps": 500,
|
| 199 |
+
"stateful_callbacks": {
|
| 200 |
+
"EarlyStoppingCallback": {
|
| 201 |
+
"args": {
|
| 202 |
+
"early_stopping_patience": 3,
|
| 203 |
+
"early_stopping_threshold": 0.0
|
| 204 |
+
},
|
| 205 |
+
"attributes": {
|
| 206 |
+
"early_stopping_patience_counter": 0
|
| 207 |
+
}
|
| 208 |
+
},
|
| 209 |
+
"TrainerControl": {
|
| 210 |
+
"args": {
|
| 211 |
+
"should_epoch_stop": false,
|
| 212 |
+
"should_evaluate": false,
|
| 213 |
+
"should_log": false,
|
| 214 |
+
"should_save": true,
|
| 215 |
+
"should_training_stop": false
|
| 216 |
+
},
|
| 217 |
+
"attributes": {}
|
| 218 |
+
}
|
| 219 |
+
},
|
| 220 |
+
"total_flos": 0.0,
|
| 221 |
+
"train_batch_size": 8,
|
| 222 |
+
"trial_name": null,
|
| 223 |
+
"trial_params": null
|
| 224 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-720/trainer_state.json
ADDED
|
@@ -0,0 +1,327 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 450,
|
| 3 |
+
"best_metric": 0.8334384195039932,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-viclickbait2025-bamibert-source-source-05-s44-20260827T175045Z-fe731097/trainer/checkpoint-450",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 720,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.2801120448179272,
|
| 14 |
+
"grad_norm": 28.82010269165039,
|
| 15 |
+
"learning_rate": 2.666666666666667e-06,
|
| 16 |
+
"loss": 2.5697,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.5602240896358543,
|
| 21 |
+
"grad_norm": 7.371585845947266,
|
| 22 |
+
"learning_rate": 5.444444444444445e-06,
|
| 23 |
+
"loss": 2.5197,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.8403361344537815,
|
| 28 |
+
"grad_norm": 15.176864624023438,
|
| 29 |
+
"learning_rate": 8.222222222222222e-06,
|
| 30 |
+
"loss": 2.4669,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6656151419558359,
|
| 36 |
+
"eval_loss": 0.5750473737716675,
|
| 37 |
+
"eval_macro_f1": 0.45522049286640726,
|
| 38 |
+
"eval_runtime": 0.2786,
|
| 39 |
+
"eval_samples_per_second": 1137.717,
|
| 40 |
+
"eval_steps_per_second": 71.78,
|
| 41 |
+
"eval_weighted_f1": 0.554543855126899,
|
| 42 |
+
"step": 90
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.112044817927171,
|
| 46 |
+
"grad_norm": 21.1071720123291,
|
| 47 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 48 |
+
"loss": 2.2098,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.392156862745098,
|
| 53 |
+
"grad_norm": 37.07575988769531,
|
| 54 |
+
"learning_rate": 1.377777777777778e-05,
|
| 55 |
+
"loss": 1.9603,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.6722689075630253,
|
| 60 |
+
"grad_norm": 27.69922637939453,
|
| 61 |
+
"learning_rate": 1.6555555555555556e-05,
|
| 62 |
+
"loss": 2.0119,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.9523809523809523,
|
| 67 |
+
"grad_norm": 35.82430648803711,
|
| 68 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 69 |
+
"loss": 1.7122,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.0,
|
| 74 |
+
"eval_accuracy": 0.8012618296529969,
|
| 75 |
+
"eval_loss": 0.4539729058742523,
|
| 76 |
+
"eval_macro_f1": 0.7639724392232872,
|
| 77 |
+
"eval_runtime": 0.3245,
|
| 78 |
+
"eval_samples_per_second": 976.963,
|
| 79 |
+
"eval_steps_per_second": 61.638,
|
| 80 |
+
"eval_weighted_f1": 0.79149556073093,
|
| 81 |
+
"step": 180
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.224089635854342,
|
| 85 |
+
"grad_norm": 42.21854019165039,
|
| 86 |
+
"learning_rate": 1.9765432098765434e-05,
|
| 87 |
+
"loss": 1.3959,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 2.504201680672269,
|
| 92 |
+
"grad_norm": 27.365318298339844,
|
| 93 |
+
"learning_rate": 1.9456790123456793e-05,
|
| 94 |
+
"loss": 1.6084,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.784313725490196,
|
| 99 |
+
"grad_norm": 47.25198745727539,
|
| 100 |
+
"learning_rate": 1.9148148148148148e-05,
|
| 101 |
+
"loss": 1.4331,
|
| 102 |
+
"step": 250
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 3.0,
|
| 106 |
+
"eval_accuracy": 0.8422712933753943,
|
| 107 |
+
"eval_loss": 0.3598579168319702,
|
| 108 |
+
"eval_macro_f1": 0.8236537605696483,
|
| 109 |
+
"eval_runtime": 0.2716,
|
| 110 |
+
"eval_samples_per_second": 1167.18,
|
| 111 |
+
"eval_steps_per_second": 73.639,
|
| 112 |
+
"eval_weighted_f1": 0.8404637659185258,
|
| 113 |
+
"step": 270
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.0560224089635852,
|
| 117 |
+
"grad_norm": 38.71910858154297,
|
| 118 |
+
"learning_rate": 1.8839506172839507e-05,
|
| 119 |
+
"loss": 1.3697,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 3.3361344537815127,
|
| 124 |
+
"grad_norm": 30.123592376708984,
|
| 125 |
+
"learning_rate": 1.8530864197530866e-05,
|
| 126 |
+
"loss": 1.1044,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 3.61624649859944,
|
| 131 |
+
"grad_norm": 50.46794891357422,
|
| 132 |
+
"learning_rate": 1.8222222222222224e-05,
|
| 133 |
+
"loss": 0.9246,
|
| 134 |
+
"step": 325
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.896358543417367,
|
| 138 |
+
"grad_norm": 33.4803466796875,
|
| 139 |
+
"learning_rate": 1.7913580246913583e-05,
|
| 140 |
+
"loss": 1.1125,
|
| 141 |
+
"step": 350
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 4.0,
|
| 145 |
+
"eval_accuracy": 0.8201892744479495,
|
| 146 |
+
"eval_loss": 0.5016289949417114,
|
| 147 |
+
"eval_macro_f1": 0.7877057558774805,
|
| 148 |
+
"eval_runtime": 0.2859,
|
| 149 |
+
"eval_samples_per_second": 1108.765,
|
| 150 |
+
"eval_steps_per_second": 69.954,
|
| 151 |
+
"eval_weighted_f1": 0.8120683948053322,
|
| 152 |
+
"step": 360
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 4.168067226890757,
|
| 156 |
+
"grad_norm": 29.374866485595703,
|
| 157 |
+
"learning_rate": 1.760493827160494e-05,
|
| 158 |
+
"loss": 0.8453,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 4.448179271708684,
|
| 163 |
+
"grad_norm": 44.94426727294922,
|
| 164 |
+
"learning_rate": 1.7296296296296297e-05,
|
| 165 |
+
"loss": 0.8138,
|
| 166 |
+
"step": 400
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 4.728291316526611,
|
| 170 |
+
"grad_norm": 38.09920883178711,
|
| 171 |
+
"learning_rate": 1.6987654320987656e-05,
|
| 172 |
+
"loss": 0.7823,
|
| 173 |
+
"step": 425
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 5.0,
|
| 177 |
+
"grad_norm": 4.866731643676758,
|
| 178 |
+
"learning_rate": 1.6679012345679015e-05,
|
| 179 |
+
"loss": 0.767,
|
| 180 |
+
"step": 450
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 5.0,
|
| 184 |
+
"eval_accuracy": 0.8422712933753943,
|
| 185 |
+
"eval_loss": 0.3990023732185364,
|
| 186 |
+
"eval_macro_f1": 0.8334384195039932,
|
| 187 |
+
"eval_runtime": 0.2885,
|
| 188 |
+
"eval_samples_per_second": 1098.655,
|
| 189 |
+
"eval_steps_per_second": 69.316,
|
| 190 |
+
"eval_weighted_f1": 0.8446912588196138,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 5.280112044817927,
|
| 195 |
+
"grad_norm": 96.87423706054688,
|
| 196 |
+
"learning_rate": 1.6370370370370374e-05,
|
| 197 |
+
"loss": 0.6863,
|
| 198 |
+
"step": 475
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 5.560224089635854,
|
| 202 |
+
"grad_norm": 61.71138381958008,
|
| 203 |
+
"learning_rate": 1.606172839506173e-05,
|
| 204 |
+
"loss": 0.4993,
|
| 205 |
+
"step": 500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 5.840336134453781,
|
| 209 |
+
"grad_norm": 15.314275741577148,
|
| 210 |
+
"learning_rate": 1.5753086419753088e-05,
|
| 211 |
+
"loss": 0.5873,
|
| 212 |
+
"step": 525
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 6.0,
|
| 216 |
+
"eval_accuracy": 0.8454258675078864,
|
| 217 |
+
"eval_loss": 0.5390811562538147,
|
| 218 |
+
"eval_macro_f1": 0.820562583030093,
|
| 219 |
+
"eval_runtime": 0.2819,
|
| 220 |
+
"eval_samples_per_second": 1124.693,
|
| 221 |
+
"eval_steps_per_second": 70.959,
|
| 222 |
+
"eval_weighted_f1": 0.8401582224914047,
|
| 223 |
+
"step": 540
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 6.1120448179271705,
|
| 227 |
+
"grad_norm": 51.439117431640625,
|
| 228 |
+
"learning_rate": 1.5444444444444446e-05,
|
| 229 |
+
"loss": 0.3745,
|
| 230 |
+
"step": 550
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 6.392156862745098,
|
| 234 |
+
"grad_norm": 13.093758583068848,
|
| 235 |
+
"learning_rate": 1.5135802469135803e-05,
|
| 236 |
+
"loss": 0.3803,
|
| 237 |
+
"step": 575
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 6.6722689075630255,
|
| 241 |
+
"grad_norm": 81.52074432373047,
|
| 242 |
+
"learning_rate": 1.4827160493827162e-05,
|
| 243 |
+
"loss": 0.5265,
|
| 244 |
+
"step": 600
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 6.9523809523809526,
|
| 248 |
+
"grad_norm": 2.439595937728882,
|
| 249 |
+
"learning_rate": 1.4518518518518521e-05,
|
| 250 |
+
"loss": 0.4071,
|
| 251 |
+
"step": 625
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 7.0,
|
| 255 |
+
"eval_accuracy": 0.8485804416403786,
|
| 256 |
+
"eval_loss": 0.6375619173049927,
|
| 257 |
+
"eval_macro_f1": 0.8307076101468625,
|
| 258 |
+
"eval_runtime": 0.2907,
|
| 259 |
+
"eval_samples_per_second": 1090.615,
|
| 260 |
+
"eval_steps_per_second": 68.809,
|
| 261 |
+
"eval_weighted_f1": 0.8468452152817847,
|
| 262 |
+
"step": 630
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 7.224089635854342,
|
| 266 |
+
"grad_norm": 33.991580963134766,
|
| 267 |
+
"learning_rate": 1.4209876543209878e-05,
|
| 268 |
+
"loss": 0.3443,
|
| 269 |
+
"step": 650
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 7.504201680672269,
|
| 273 |
+
"grad_norm": 229.5233917236328,
|
| 274 |
+
"learning_rate": 1.3901234567901237e-05,
|
| 275 |
+
"loss": 0.4044,
|
| 276 |
+
"step": 675
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 7.784313725490196,
|
| 280 |
+
"grad_norm": 351.1485290527344,
|
| 281 |
+
"learning_rate": 1.3592592592592592e-05,
|
| 282 |
+
"loss": 0.4095,
|
| 283 |
+
"step": 700
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 8.0,
|
| 287 |
+
"eval_accuracy": 0.8233438485804416,
|
| 288 |
+
"eval_loss": 0.9755781292915344,
|
| 289 |
+
"eval_macro_f1": 0.8134510298444724,
|
| 290 |
+
"eval_runtime": 0.2965,
|
| 291 |
+
"eval_samples_per_second": 1069.09,
|
| 292 |
+
"eval_steps_per_second": 67.45,
|
| 293 |
+
"eval_weighted_f1": 0.8260542098779673,
|
| 294 |
+
"step": 720
|
| 295 |
+
}
|
| 296 |
+
],
|
| 297 |
+
"logging_steps": 25,
|
| 298 |
+
"max_steps": 1800,
|
| 299 |
+
"num_input_tokens_seen": 0,
|
| 300 |
+
"num_train_epochs": 20,
|
| 301 |
+
"save_steps": 500,
|
| 302 |
+
"stateful_callbacks": {
|
| 303 |
+
"EarlyStoppingCallback": {
|
| 304 |
+
"args": {
|
| 305 |
+
"early_stopping_patience": 3,
|
| 306 |
+
"early_stopping_threshold": 0.0
|
| 307 |
+
},
|
| 308 |
+
"attributes": {
|
| 309 |
+
"early_stopping_patience_counter": 3
|
| 310 |
+
}
|
| 311 |
+
},
|
| 312 |
+
"TrainerControl": {
|
| 313 |
+
"args": {
|
| 314 |
+
"should_epoch_stop": false,
|
| 315 |
+
"should_evaluate": false,
|
| 316 |
+
"should_log": false,
|
| 317 |
+
"should_save": true,
|
| 318 |
+
"should_training_stop": true
|
| 319 |
+
},
|
| 320 |
+
"attributes": {}
|
| 321 |
+
}
|
| 322 |
+
},
|
| 323 |
+
"total_flos": 0.0,
|
| 324 |
+
"train_batch_size": 8,
|
| 325 |
+
"trial_name": null,
|
| 326 |
+
"trial_params": null
|
| 327 |
+
}
|