Update README.md
Browse files
README.md
CHANGED
|
@@ -10,6 +10,31 @@ base_model: openai/gpt-oss-120b
|
|
| 10 |
|
| 11 |
# gpt-oss-120b-qx64-mlx
|
| 12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
This model [gpt-oss-120b-qx64-mlx](https://huggingface.co/gpt-oss-120b-qx64-mlx) was
|
| 14 |
converted to MLX format from [openai/gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b)
|
| 15 |
using mlx-lm version **0.27.1**.
|
|
|
|
| 10 |
|
| 11 |
# gpt-oss-120b-qx64-mlx
|
| 12 |
|
| 13 |
+
The reason I created the qx64 and qx65 quants: I was looking to write some Perl as a Postgres function.
|
| 14 |
+
|
| 15 |
+
Most other quants simplify, offer really well written PL/PGSQL instead.
|
| 16 |
+
|
| 17 |
+
But I wanted PL/Perl. I am that guy.
|
| 18 |
+
|
| 19 |
+
The [qx65 quant](https://huggingface.co/nightmedia/gpt-oss-120b-qx65-mlx) has given me what I asked.
|
| 20 |
+
|
| 21 |
+
It followed instructions.
|
| 22 |
+
|
| 23 |
+
Then I asked the qx64 the same question--why did you follow my instructions: I showed it this prompt.
|
| 24 |
+
|
| 25 |
+
The qx65 gave me a very short, clean answer I could put as a comment in the code.
|
| 26 |
+
|
| 27 |
+
The qx64 gave me the history of PL/Perl and how many nice things I could do with it.
|
| 28 |
+
|
| 29 |
+
Until the performance metrics are available, please use these models with caution.
|
| 30 |
+
|
| 31 |
+
-G
|
| 32 |
+
|
| 33 |
+
```bash
|
| 34 |
+
75.26 tok/sec
|
| 35 |
+
9338 tokens
|
| 36 |
+
2.58s to first token
|
| 37 |
+
```
|
| 38 |
This model [gpt-oss-120b-qx64-mlx](https://huggingface.co/gpt-oss-120b-qx64-mlx) was
|
| 39 |
converted to MLX format from [openai/gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b)
|
| 40 |
using mlx-lm version **0.27.1**.
|