nemotron-3-nano:4b-mxfp8

918.6K 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.

tools thinking cloud 4b 30b
... /
json
7c4c0f4472e2 ยท 2.5kB
{
"architectures": [
"NemotronHForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"head_dim": 128,
"auto_map": {
"AutoConfig": "configuration_nemotron_h.NemotronHConfig",
"AutoModelForCausalLM": "modeling_nemotron_h.NemotronHForCausalLM"
},
"bos_token_id": 1,
"chunk_size": 256,
"conv_kernel": 4,
"eos_token_id": 2,
"expand": 2,
"hidden_dropout": 0.0,
"hidden_size": 3136,
"hybrid_override_pattern": "M-M-M-MM-M-M*-M-M*-M-M-M*-M-M-MM*-MMM-M-M-",
"initializer_range": 0.02,
"intermediate_size": 12544,
"layer_norm_epsilon": 1e-05,
"mamba_head_dim": 80,
"mamba_hidden_act": "silu",
"mamba_num_heads": 96,
"mamba_proj_bias": false,
"max_position_embeddings": 262144,
"mlp_bias": false,
"mlp_hidden_act": "relu2",
"model_type": "nemotron_h",
"n_groups": 8,
"num_attention_heads": 40,
"num_hidden_layers": 42,
"num_key_value_heads": 8,
"num_logits_to_keep": 1,
"pad_token_id": 0,
"rescale_prenorm_residual": true,
"residual_in_fp32": false,
"rms_norm_eps": 1e-05,
"sliding_window": null,
"ssm_state_size": 128,
"tie_word_embeddings": false,
"time_step_floor": 0.0001,
"time_step_max": 0.1,
"time_step_min": 0.001,
"time_step_rank": 256,
"torch_dtype": "bfloat16",
"transformers_version": "4.53.0",
"use_bias": false,
"use_cache": true,
"use_conv_bias": true,
"use_mamba_kernels": true,
"vocab_size": 131072,
"quantization_config": {
"ignore": [
"backbone.layers.0.mixer.conv1d",
"backbone.layers.11.mixer.conv1d",
"backbone.layers.14.mixer.conv1d",
"backbone.layers.16.mixer.conv1d",
"backbone.layers.19.mixer.conv1d",
"backbone.layers.2.mixer.conv1d",
"backbone.layers.21.mixer.conv1d",
"backbone.layers.23.mixer.conv1d",
"backbone.layers.26.mixer.conv1d",
"backbone.layers.28.mixer.conv1d",
"backbone.layers.30.mixer.conv1d",
"backbone.layers.31.mixer.conv1d",
"backbone.layers.34.mixer.conv1d",
"backbone.layers.35.mixer.conv1d",
"backbone.layers.36.mixer.conv1d",
"backbone.layers.38.mixer.conv1d",
"backbone.layers.4.mixer.conv1d",
"backbone.layers.40.mixer.conv1d",
"backbone.layers.6.mixer.conv1d",
"backbone.layers.7.mixer.conv1d",
"backbone.layers.9.mixer.conv1d"
],
"quant_algo": "MXFP8",
"producer": {
"name": "modelopt",
"version": "0.0.0"
},
"quant_method": "modelopt"
},
"quantization": {
"bits": 8,
"mode": "mxfp8"
}
}