BrokeC commited on
Commit
d8a4531
·
verified ·
1 Parent(s): 46135ef

Upload processor_config.json

Browse files

## Fix: Add missing `spatial_merge_size` to processor config

### Problem

Missing `spatial_merge_size` parameter causes vision inference to crash:
ValueError: Number of image token positions (5476) does not match number of image features (1369)

### Solution

Add `"spatial_merge_size": 2` to `processor_config.json`:

```json
{
"image_break_token": "[IMG_BREAK]",
"image_end_token": "[IMG_END]",
"image_token": "[IMG]",
"patch_size": 14,
"processor_class": "PixtralProcessor",
"spatial_merge_size": 2
}

This matches the original https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/blob/main/processor_config.json and config.json.

✅ Tested on M2 Max 64GB - vision inference works after fix.

Same fixes needed for: -4bit, -6bit, -8bit, -bf16 (all 0.1.19 conversions)
```

Files changed (1) hide show
  1. processor_config.json +2 -1
processor_config.json CHANGED
@@ -3,5 +3,6 @@
3
  "image_end_token": "[IMG_END]",
4
  "image_token": "[IMG]",
5
  "patch_size": 14,
6
- "processor_class": "PixtralProcessor"
 
7
  }
 
3
  "image_end_token": "[IMG_END]",
4
  "image_token": "[IMG]",
5
  "patch_size": 14,
6
+ "processor_class": "PixtralProcessor",
7
+ "spatial_merge_size": 2
8
  }