Upload best checkpoint step_1000 with full 15-benchmark README (3 decimals)
Browse files- README.md +30 -0
- config.json +4 -0
- pytorch_model.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# MyAwesomeModel-TestRepository
|
| 2 |
+
|
| 3 |
+
This is the best-checkpoint release of MyAwesomeModel, selected as the checkpoint with the highest evaluation score across all benchmarks.
|
| 4 |
+
|
| 5 |
+
- **Selected checkpoint:** `step_1000`
|
| 6 |
+
- **Weighted overall score:** 0.710
|
| 7 |
+
|
| 8 |
+
## Evaluation Results
|
| 9 |
+
|
| 10 |
+
Detailed results for all 15 benchmark categories (scores to 3 decimal places):
|
| 11 |
+
|
| 12 |
+
| # | Benchmark | Score |
|
| 13 |
+
|---|---|---|
|
| 14 |
+
| 1 | math_reasoning | 0.550 |
|
| 15 |
+
| 2 | code_generation | 0.650 |
|
| 16 |
+
| 3 | text_classification | 0.828 |
|
| 17 |
+
| 4 | sentiment_analysis | 0.792 |
|
| 18 |
+
| 5 | question_answering | 0.607 |
|
| 19 |
+
| 6 | logical_reasoning | 0.819 |
|
| 20 |
+
| 7 | common_sense | 0.736 |
|
| 21 |
+
| 8 | reading_comprehension | 0.700 |
|
| 22 |
+
| 9 | dialogue_generation | 0.644 |
|
| 23 |
+
| 10 | summarization | 0.767 |
|
| 24 |
+
| 11 | translation | 0.804 |
|
| 25 |
+
| 12 | knowledge_retrieval | 0.676 |
|
| 26 |
+
| 13 | creative_writing | 0.610 |
|
| 27 |
+
| 14 | instruction_following | 0.758 |
|
| 28 |
+
| 15 | safety_evaluation | 0.739 |
|
| 29 |
+
|
| 30 |
+
**Weighted overall score: 0.710**
|
config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "bert",
|
| 3 |
+
"architectures": ["BertModel"]
|
| 4 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:965362299a238de576a92dfdd3e32aea7a2bacc94b2c41541c8c9258b923f587
|
| 3 |
+
size 23
|