Update README.md
Browse files
README.md
CHANGED
|
@@ -4,7 +4,6 @@ pipeline_tag: text-generation
|
|
| 4 |
inference: false
|
| 5 |
license: apache-2.0
|
| 6 |
base_model: openai/gpt-oss-120b
|
| 7 |
-
base_model_relation: adapter
|
| 8 |
language:
|
| 9 |
- en
|
| 10 |
- ro
|
|
@@ -71,7 +70,7 @@ Perplexity (PPL) streaming evaluation on WikiText-2 (raw, test); fast preset wit
|
|
| 71 |
| MLX 8-bit (gs=32) | 7.39 |
|
| 72 |
| MLX 6-bit (gs=64) | 7.40 |
|
| 73 |
|
| 74 |
-
Notes
|
| 75 |
|
| 76 |
- Results from local runs on Apple Silicon using MLX; numbers vary slightly with tokenizer details, logits dtype, and token subset.
|
| 77 |
- For more sensitive comparisons, use overlapping windows (e.g., `--stride 512`) and evaluate the full split.
|
|
|
|
| 4 |
inference: false
|
| 5 |
license: apache-2.0
|
| 6 |
base_model: openai/gpt-oss-120b
|
|
|
|
| 7 |
language:
|
| 8 |
- en
|
| 9 |
- ro
|
|
|
|
| 70 |
| MLX 8-bit (gs=32) | 7.39 |
|
| 71 |
| MLX 6-bit (gs=64) | 7.40 |
|
| 72 |
|
| 73 |
+
### Notes
|
| 74 |
|
| 75 |
- Results from local runs on Apple Silicon using MLX; numbers vary slightly with tokenizer details, logits dtype, and token subset.
|
| 76 |
- For more sensitive comparisons, use overlapping windows (e.g., `--stride 512`) and evaluate the full split.
|