Instructions to use janhq/Jan-v2-VL-high-4bit-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use janhq/Jan-v2-VL-high-4bit-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Jan-v2-VL-high-4bit-mlx janhq/Jan-v2-VL-high-4bit-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Update config.json
Browse files- config.json +3 -2
config.json
CHANGED
|
@@ -3,7 +3,7 @@
|
|
| 3 |
"Qwen3VLForConditionalGeneration"
|
| 4 |
],
|
| 5 |
"dtype": "float32",
|
| 6 |
-
"eos_token_id":
|
| 7 |
"image_token_id": 151655,
|
| 8 |
"model_type": "qwen3_vl",
|
| 9 |
"pad_token_id": 151643,
|
|
@@ -45,7 +45,8 @@
|
|
| 45 |
},
|
| 46 |
"rope_theta": 5000000,
|
| 47 |
"use_cache": true,
|
| 48 |
-
"vocab_size": 151936
|
|
|
|
| 49 |
},
|
| 50 |
"tie_word_embeddings": false,
|
| 51 |
"transformers_version": "4.57.0",
|
|
|
|
| 3 |
"Qwen3VLForConditionalGeneration"
|
| 4 |
],
|
| 5 |
"dtype": "float32",
|
| 6 |
+
"eos_token_id": null,
|
| 7 |
"image_token_id": 151655,
|
| 8 |
"model_type": "qwen3_vl",
|
| 9 |
"pad_token_id": 151643,
|
|
|
|
| 45 |
},
|
| 46 |
"rope_theta": 5000000,
|
| 47 |
"use_cache": true,
|
| 48 |
+
"vocab_size": 151936,
|
| 49 |
+
"tie_word_embeddings": false
|
| 50 |
},
|
| 51 |
"tie_word_embeddings": false,
|
| 52 |
"transformers_version": "4.57.0",
|