GitHub
183.6K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
nemotron-3-nano
:4b-mlx
918.6K
Downloads
Updated
an hour ago
Nemotron-3-Nano is a new Standard for Efficient, Open, and Intelligent Agentic Models, now updated with a 4B parameter count model.
Nemotron-3-Nano is a new Standard for Efficient, Open, and Intelligent Agentic Models, now updated with a 4B parameter count model.
Cancel
tools
thinking
cloud
4b
30b
nemotron-3-nano:4b-mlx
...
/
json
762b1e4b79b3 ยท 13kB
{
"producer": {
"name": "modelopt",
"version": "0.0.0"
},
"quantization": {
"quant_algo": "MIXED_PRECISION",
"kv_cache_quant_algo": "NVFP4",
"quantized_layers": {
"backbone.embeddings": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.0.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.0.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.1.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.1.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.2.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.2.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.3.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.3.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.4.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.4.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.5.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.5.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.6.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.6.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.7.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.7.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.8.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.8.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.9.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.9.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.10.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.10.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.11.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.11.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.12.mixer.q_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.12.mixer.k_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.12.mixer.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.12.mixer.o_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.13.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.13.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.14.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.14.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.15.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.15.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.16.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.16.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.17.mixer.q_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.17.mixer.k_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.17.mixer.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.17.mixer.o_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.18.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.18.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.19.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.19.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.20.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.20.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.21.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.21.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.22.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.22.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.23.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.23.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.24.mixer.q_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.24.mixer.k_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.24.mixer.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.24.mixer.o_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.25.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.25.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.26.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.26.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.27.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.27.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.28.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.28.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.29.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.29.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.30.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.30.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.31.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.31.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.32.mixer.q_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.32.mixer.k_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.32.mixer.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.32.mixer.o_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.33.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.33.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.34.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.34.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.35.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.35.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.36.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.36.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.37.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.37.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.38.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.38.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.39.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.39.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.40.mixer.in_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"backbone.layers.40.mixer.out_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"backbone.layers.41.mixer.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"backbone.layers.41.mixer.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"lm_head": {
"quant_algo": "MXFP8",
"group_size": 32
}
}
}
}