Instructions to use qrizan/pii-masking-qwen2.5-3b-dpo_beta0_5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use qrizan/pii-masking-qwen2.5-3b-dpo_beta0_5 with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/qwen2.5-3b-instruct-unsloth-bnb-4bit") model = PeftModel.from_pretrained(base_model, "qrizan/pii-masking-qwen2.5-3b-dpo_beta0_5") - Notebooks
- Google Colab
- Kaggle
Upload training_metadata.json with huggingface_hub
Browse files- training_metadata.json +19 -1
training_metadata.json
CHANGED
|
@@ -17,7 +17,7 @@
|
|
| 17 |
"learning_rate": 5e-05,
|
| 18 |
"actual_steps": 120,
|
| 19 |
"train_loss": 0.02035108965168483,
|
| 20 |
-
"train_time_s":
|
| 21 |
},
|
| 22 |
"beta_sweep_results": {
|
| 23 |
"0.0": {
|
|
@@ -51,6 +51,12 @@
|
|
| 51 |
"false_mask_count": 7,
|
| 52 |
"mean_jaccard": 0.30612244897959184
|
| 53 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
"eval_regressed_nonholdout": false
|
| 55 |
},
|
| 56 |
"0.1": {
|
|
@@ -84,6 +90,12 @@
|
|
| 84 |
"false_mask_count": 0,
|
| 85 |
"mean_jaccard": 0.4238095238095238
|
| 86 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
"eval_regressed_nonholdout": true
|
| 88 |
},
|
| 89 |
"0.5": {
|
|
@@ -117,6 +129,12 @@
|
|
| 117 |
"false_mask_count": 0,
|
| 118 |
"mean_jaccard": 0.7714285714285715
|
| 119 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 120 |
"eval_regressed_nonholdout": false
|
| 121 |
}
|
| 122 |
},
|
|
|
|
| 17 |
"learning_rate": 5e-05,
|
| 18 |
"actual_steps": 120,
|
| 19 |
"train_loss": 0.02035108965168483,
|
| 20 |
+
"train_time_s": 418.0
|
| 21 |
},
|
| 22 |
"beta_sweep_results": {
|
| 23 |
"0.0": {
|
|
|
|
| 51 |
"false_mask_count": 7,
|
| 52 |
"mean_jaccard": 0.30612244897959184
|
| 53 |
},
|
| 54 |
+
"eval_exact_match": {
|
| 55 |
+
"all": 0.93,
|
| 56 |
+
"holdout": 0.5333333333333333,
|
| 57 |
+
"non_holdout": 1.0,
|
| 58 |
+
"t8": 0.0
|
| 59 |
+
},
|
| 60 |
"eval_regressed_nonholdout": false
|
| 61 |
},
|
| 62 |
"0.1": {
|
|
|
|
| 90 |
"false_mask_count": 0,
|
| 91 |
"mean_jaccard": 0.4238095238095238
|
| 92 |
},
|
| 93 |
+
"eval_exact_match": {
|
| 94 |
+
"all": 0.76,
|
| 95 |
+
"holdout": 0.13333333333333333,
|
| 96 |
+
"non_holdout": 0.8705882352941177,
|
| 97 |
+
"t8": 0.0
|
| 98 |
+
},
|
| 99 |
"eval_regressed_nonholdout": true
|
| 100 |
},
|
| 101 |
"0.5": {
|
|
|
|
| 129 |
"false_mask_count": 0,
|
| 130 |
"mean_jaccard": 0.7714285714285715
|
| 131 |
},
|
| 132 |
+
"eval_exact_match": {
|
| 133 |
+
"all": 0.97,
|
| 134 |
+
"holdout": 0.8,
|
| 135 |
+
"non_holdout": 1.0,
|
| 136 |
+
"t8": 0.5714285714285714
|
| 137 |
+
},
|
| 138 |
"eval_regressed_nonholdout": false
|
| 139 |
}
|
| 140 |
},
|