Keep original README content and update only benchmark scores
Browse files
README.md
CHANGED
|
@@ -59,39 +59,6 @@ Beyond its improved reasoning capabilities, this version also offers a reduced h
|
|
| 59 |
### Overall Performance Summary
|
| 60 |
The MyAwesomeModel demonstrates strong performance across all evaluated benchmark categories, with particularly notable results in reasoning and generation tasks.
|
| 61 |
|
| 62 |
-
|
| 63 |
-
## Selected Checkpoint
|
| 64 |
-
|
| 65 |
-
This repository contains `step_1000`, selected after scanning all checkpoint directories in the workspace and choosing the checkpoint with the highest `eval_accuracy`.
|
| 66 |
-
|
| 67 |
-
- Selection metric: `eval_accuracy`
|
| 68 |
-
- Best checkpoint: `step_1000`
|
| 69 |
-
- Best `eval_accuracy`: **0.710**
|
| 70 |
-
|
| 71 |
-
### Detailed Evaluation Results for Selected Checkpoint
|
| 72 |
-
|
| 73 |
-
The selected checkpoint is `step_1000`, chosen by highest `eval_accuracy` across the checkpoints found in the workspace. All benchmark scores are reported to **three decimal places**.
|
| 74 |
-
|
| 75 |
-
- Overall selection metric, `eval_accuracy`: **0.710**
|
| 76 |
-
|
| 77 |
-
| Category | Benchmark | Score |
|
| 78 |
-
|---|---|---:|
|
| 79 |
-
| Core Reasoning Tasks | Math Reasoning | 0.550 |
|
| 80 |
-
| Core Reasoning Tasks | Logical Reasoning | 0.819 |
|
| 81 |
-
| Core Reasoning Tasks | Common Sense | 0.736 |
|
| 82 |
-
| Language Understanding | Reading Comprehension | 0.700 |
|
| 83 |
-
| Language Understanding | Question Answering | 0.607 |
|
| 84 |
-
| Language Understanding | Text Classification | 0.828 |
|
| 85 |
-
| Language Understanding | Sentiment Analysis | 0.792 |
|
| 86 |
-
| Generation Tasks | Code Generation | 0.650 |
|
| 87 |
-
| Generation Tasks | Creative Writing | 0.610 |
|
| 88 |
-
| Generation Tasks | Dialogue Generation | 0.644 |
|
| 89 |
-
| Generation Tasks | Summarization | 0.767 |
|
| 90 |
-
| Specialized Capabilities | Translation | 0.804 |
|
| 91 |
-
| Specialized Capabilities | Knowledge Retrieval | 0.676 |
|
| 92 |
-
| Specialized Capabilities | Instruction Following | 0.758 |
|
| 93 |
-
| Specialized Capabilities | Safety Evaluation | 0.739 |
|
| 94 |
-
|
| 95 |
## 3. Chat Website & API Platform
|
| 96 |
We offer a chat interface and API for you to interact with MyAwesomeModel. Please check our official website for more details.
|
| 97 |
|
|
|
|
| 59 |
### Overall Performance Summary
|
| 60 |
The MyAwesomeModel demonstrates strong performance across all evaluated benchmark categories, with particularly notable results in reasoning and generation tasks.
|
| 61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
## 3. Chat Website & API Platform
|
| 63 |
We offer a chat interface and API for you to interact with MyAwesomeModel. Please check our official website for more details.
|
| 64 |
|