Any-to-Any
MLX
Safetensors
English
multilingual
nemotron_h
nemotron
nemotron-h
jangtq
crack
abliterated
uncensored
multimodal
vision
audio
speech
mamba-2
Mixture of Experts
reasoning
thinking
harmbench
radio-vit
parakeet
custom_code
Instructions to use dealignai/Nemotron-3-Nano-Omni-30B-A3B-JANGTQ4-CRACK with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use dealignai/Nemotron-3-Nano-Omni-30B-A3B-JANGTQ4-CRACK with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Nemotron-3-Nano-Omni-30B-A3B-JANGTQ4-CRACK dealignai/Nemotron-3-Nano-Omni-30B-A3B-JANGTQ4-CRACK
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
| { | |
| "vectors": "vectors/jangtq4_1000pair.safetensors", | |
| "direction_layer": 42, | |
| "per_layer": true, | |
| "target_layers": [ | |
| 26, | |
| 33, | |
| 42 | |
| ], | |
| "mamba_layers": [ | |
| 28, | |
| 30, | |
| 32, | |
| 35, | |
| 37, | |
| 39, | |
| 41, | |
| 44, | |
| 46 | |
| ], | |
| "strength": 8.0, | |
| "mamba_strength": 5.0, | |
| "ablation": "mpoa", | |
| "mode": "affine", | |
| "bits": 8, | |
| "group_size": 64, | |
| "results": [ | |
| { | |
| "layer": 26, | |
| "delta_norm": 20.723915100097656, | |
| "change_pct": 25.992341788637116, | |
| "shards": [ | |
| "model-00009-of-00021.safetensors" | |
| ], | |
| "kind": "attn", | |
| "tensor": "o_proj" | |
| }, | |
| { | |
| "layer": 33, | |
| "delta_norm": 18.07789421081543, | |
| "change_pct": 20.49878806995057, | |
| "shards": [ | |
| "model-00011-of-00021.safetensors" | |
| ], | |
| "kind": "attn", | |
| "tensor": "o_proj" | |
| }, | |
| { | |
| "layer": 42, | |
| "delta_norm": 15.052383422851562, | |
| "change_pct": 15.51020741873604, | |
| "shards": [ | |
| "model-00014-of-00021.safetensors" | |
| ], | |
| "kind": "attn", | |
| "tensor": "o_proj" | |
| }, | |
| { | |
| "layer": 28, | |
| "delta_norm": 9.787358283996582, | |
| "change_pct": 13.004941193597084, | |
| "shards": [ | |
| "model-00009-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 30, | |
| "delta_norm": 8.532942771911621, | |
| "change_pct": 11.340357951919545, | |
| "shards": [ | |
| "model-00010-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 32, | |
| "delta_norm": 8.56724739074707, | |
| "change_pct": 11.437007739685978, | |
| "shards": [ | |
| "model-00011-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 35, | |
| "delta_norm": 8.362507820129395, | |
| "change_pct": 11.077500263523666, | |
| "shards": [ | |
| "model-00012-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 37, | |
| "delta_norm": 7.6969475746154785, | |
| "change_pct": 10.24664891780168, | |
| "shards": [ | |
| "model-00012-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 39, | |
| "delta_norm": 7.674215316772461, | |
| "change_pct": 10.256496862100034, | |
| "shards": [ | |
| "model-00013-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 41, | |
| "delta_norm": 7.842741966247559, | |
| "change_pct": 10.339131144518674, | |
| "shards": [ | |
| "model-00014-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 44, | |
| "delta_norm": 7.9808454513549805, | |
| "change_pct": 10.658205724749374, | |
| "shards": [ | |
| "model-00014-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| }, | |
| { | |
| "layer": 46, | |
| "delta_norm": 8.239596366882324, | |
| "change_pct": 11.111798608983312, | |
| "shards": [ | |
| "model-00015-of-00021.safetensors" | |
| ], | |
| "kind": "mamba", | |
| "tensor": "out_proj" | |
| } | |
| ] | |
| } |