gemma4:26b-nvfp4

25.5M 51 minutes ago

Gemma 4 models are designed to deliver frontier-level performance at each size. They are well-suited for reasoning, agentic workflows, coding, and multimodal understanding.

vision tools thinking audio cloud e2b e4b 12b 26b 31b
38a8b3f0379d · 29kB
{
"producer": {
"name": "modelopt",
"version": "0.47.0rc1.dev36+g613e5e8b2"
},
"quantization": {
"kv_cache_quant_algo": "NVFP4",
"quant_algo": "MIXED_PRECISION",
"quantized_layers": {
"model.language_model.embed_tokens": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.0.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.0.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.0.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.0.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.0.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.0.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.0.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.0.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.1.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.1.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.1.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.1.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.1.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.1.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.1.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.1.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.10.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.10.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.11.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.11.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.12.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.13.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.14.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.14.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.14.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.14.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.14.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.14.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.14.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.14.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.15.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.15.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.16.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.17.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.17.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.18.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.19.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.2.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.2.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.2.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.2.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.2.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.2.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.2.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.2.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.20.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.20.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.20.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.20.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.20.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.20.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.20.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.20.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.21.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.21.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.22.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.23.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.23.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.24.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.25.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.26.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.26.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.26.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.26.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.26.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.26.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.26.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.26.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.27.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.27.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.27.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.27.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.27.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.27.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.27.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.27.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.28.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.28.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.28.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.28.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.28.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.28.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.28.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.28.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.29.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.29.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.29.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.29.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.29.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.29.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.29.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.3.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.4.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.5.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.5.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.6.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.7.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.8.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.8.mlp.down_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.8.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.8.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.8.self_attn.k_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.8.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.8.self_attn.q_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.8.self_attn.v_proj": {
"group_size": 32,
"quant_algo": "MXFP8"
},
"model.language_model.layers.9.experts": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.mlp.down_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.mlp.gate_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.mlp.up_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.self_attn.k_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.self_attn.o_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.self_attn.q_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
},
"model.language_model.layers.9.self_attn.v_proj": {
"group_size": 16,
"quant_algo": "NVFP4"
}
}
}
}