GitHub
187K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
decadence
/
gemma-4
:12b-it-abliterated-mlx-8bit
654
Downloads
Updated
1 week ago
12B — QAT — GGUF — MLX — uncensored — sampling | ~7-13 GB
12B — QAT — GGUF — MLX — uncensored — sampling | ~7-13 GB
Cancel
vision
tools
thinking
audio
gemma-4:12b-it-abliterated-mlx-8bit
...
/
json
68a02465e5a8 · 104kB
{
"architectures": [
"Gemma4UnifiedForConditionalGeneration"
],
"audio_config": {
"_name_or_path": "",
"architectures": null,
"audio_embed_dim": 640,
"chunk_size_feed_forward": 0,
"dtype": "bfloat16",
"id2label": {
"0": "LABEL_0",
"1": "LABEL_1"
},
"initializer_range": 0.02,
"is_encoder_decoder": false,
"label2id": {
"LABEL_0": 0,
"LABEL_1": 1
},
"model_type": "gemma4_unified_audio",
"output_attentions": false,
"output_hidden_states": false,
"problem_type": null,
"return_dict": true,
"rms_norm_eps": 1e-06
},
"audio_token_id": 258881,
"boa_token_id": 256000,
"boi_token_id": 255999,
"dtype": "bfloat16",
"eoa_token_index": 258883,
"eoi_token_id": 258882,
"eos_token_id": [
1,
106,
50
],
"image_token_id": 258880,
"initializer_range": 0.02,
"model_type": "gemma4_unified",
"quantization": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8",
"language_model.model.layers.0.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
}
},
"quantization_config": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8",
"language_model.model.layers.0.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.0.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.1.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.2.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.3.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.4.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.5.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.6.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.7.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.8.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.9.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.10.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.11.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.12.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.13.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.14.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.15.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.16.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.17.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.18.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.19.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.20.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.21.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.22.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.23.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.24.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.25.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.26.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.27.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.28.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.29.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.30.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.31.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.32.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.33.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.34.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.35.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.36.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.37.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.38.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.39.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.40.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.41.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.42.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.43.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.44.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.45.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.v_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.46.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.q_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.k_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.self_attn.o_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.gate_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.down_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
},
"language_model.model.layers.47.mlp.up_proj": {
"group_size": 32,
"bits": 8,
"mode": "mxfp8"
}
},
"text_config": {
"attention_bias": false,
"attention_dropout": 0.0,
"attention_k_eq_v": true,
"bos_token_id": 2,
"dtype": "bfloat16",
"enable_moe_block": false,
"eos_token_id": 1,
"final_logit_softcapping": 30.0,
"global_head_dim": 512,
"head_dim": 256,
"hidden_activation": "gelu_pytorch_tanh",
"hidden_size": 3840,
"hidden_size_per_layer_input": 0,
"initializer_range": 0.02,
"intermediate_size": 15360,
"layer_types": [
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"slid