muse-glimmer:30b-nvfp4

76K 18 minutes ago

Meta's latest open model built for always-on local agents. 30B parameters, licensed under Apache 2.0 and runs on a single GPU — tuned for tool use, long tasks, and failure recovery.

vision tools thinking 30b
8fa7e6bca5a5 · 104kB
{
"producer": {
"name": "modelopt",
"version": "0.45.0.dev385+ga23390dbb.d20260731"
},
"quantization": {
"quant_algo": "MIXED_PRECISION",
"kv_cache_quant_algo": null,
"quantized_layers": {
"model.embed_tokens": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.0.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.0.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.0.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.0.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.0.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.0.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.1.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.1.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.1.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.1.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.1.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.1.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.2.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.2.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.2.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.2.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.2.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.2.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.3.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.3.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.3.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.3.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.3.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.3.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.4.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.4.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.4.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.4.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.4.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.4.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.5.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.5.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.5.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.5.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.5.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.5.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.6.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.6.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.6.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.6.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.6.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.6.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.7.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.7.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.7.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.7.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.7.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.7.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.8.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.8.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.8.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.8.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.8.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.8.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.9.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.9.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.9.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.9.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.9.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.9.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.10.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.10.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.10.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.10.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.10.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.10.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.11.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.11.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.11.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.11.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.11.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.11.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.12.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.12.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.12.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.12.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.12.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.12.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.13.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.13.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.13.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.13.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.13.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.13.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.14.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.14.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.14.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.14.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.14.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.14.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.15.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.15.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.15.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.15.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.15.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.15.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.16.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.16.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.16.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.16.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.16.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.16.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.17.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.17.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.17.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.17.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.17.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.17.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.18.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.18.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.18.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.18.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.18.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.18.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.19.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.19.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.19.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.19.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.19.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.19.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.20.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.20.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.20.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.20.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.20.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.20.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.21.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.21.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.21.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.21.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.21.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.21.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.22.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.22.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.22.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.22.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.22.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.22.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.23.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.23.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.23.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.23.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.23.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.23.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.24.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.24.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.24.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.24.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.24.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.24.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.25.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.25.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.25.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.25.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.25.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.25.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.26.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.26.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.26.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.26.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.26.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.26.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.27.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.27.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.27.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.27.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.27.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.27.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.28.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.28.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.28.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.28.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.28.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.28.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.29.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.29.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.29.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.29.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.29.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.29.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.30.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.30.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.30.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.30.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.30.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.30.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.31.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.31.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.31.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.31.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.31.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.31.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.32.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.32.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.32.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.32.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.32.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.32.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.33.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.33.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.33.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.33.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.33.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.33.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.34.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.34.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.34.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.34.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.34.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.34.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.35.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.35.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.35.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.35.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.35.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.35.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.36.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.36.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.36.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.36.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.36.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.36.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.37.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.37.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.37.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.37.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.37.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.37.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.38.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.38.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.38.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.38.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.38.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.38.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.39.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.39.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.39.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.39.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.39.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.39.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.40.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.40.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.40.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.40.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.40.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.40.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.41.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.41.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.41.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.41.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.41.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.41.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.42.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.42.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.42.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.42.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.42.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.42.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.43.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.43.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.43.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.43.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.43.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.43.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.44.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.44.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.44.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.44.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.44.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.44.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.45.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.45.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.45.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.45.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.45.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.45.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.46.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.46.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.46.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.46.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.46.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.46.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.47.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.47.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.47.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.47.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.47.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.47.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.48.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.48.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.48.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.48.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.48.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.48.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.49.attn.q_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.49.attn.k_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.49.attn.v_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.vision_encoder.transformer.49.attn.o_proj": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.49.mlp.c_fc": {
"quant_algo": "W4A16_NVFP4",
"group_size": 16
},
"model.vision_encoder.transformer.49.mlp.c_proj": {
"quant_algo": "MXFP8",
"group_size": 32
},
"model.layers.0.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.0.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.1.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.2.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.3.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.4.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.5.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.6.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.7.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.8.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.9.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.10.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.11.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.12.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.13.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.14.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.15.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.16.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.17.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.18.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.19.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.20.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.21.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.22.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.23.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.24.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.25.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.26.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.27.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.28.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.29.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.30.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.31.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.32.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.33.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.34.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.35.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.36.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.37.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.38.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.39.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.40.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.41.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.42.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.43.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.44.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.45.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.46.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.47.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.48.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.mlp.up_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.49.mlp.down_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.self_attn.q_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.self_attn.k_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.self_attn.v_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.self_attn.o_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.self_attn.output_gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16
},
"model.layers.50.mlp.gate_proj": {
"quant_algo": "NVFP4",
"group_size": 16