| { |
| "created_at": "2026-08-02T19:21:16.489151Z", |
| "plan_sha256": "3b4abbf474c75fb70f63b6db16bdedeff89122e7a6731f49cb7e1df924fd6f9d", |
| "records": [ |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.embed_tokens", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.0.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.1.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.10.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.11.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.12.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.13.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.14.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.15.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.16.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.17.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.18.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.19.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.2.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.20.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.21.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.3.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.4.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 6, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.5.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.6.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.7.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.8.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.9.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.22.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.23.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.24.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.25.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.26.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.27.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.28.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.29.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.30.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.31.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.32.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.33.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.34.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.mlp.down_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.mlp.gate_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.mlp.up_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.self_attn.k_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.self_attn.o_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 64, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.self_attn.q_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| }, |
| { |
| "bits": 8, |
| "fallback": false, |
| "group_size": 32, |
| "metadata": {}, |
| "method": "affine", |
| "module_path": "model.layers.35.self_attn.v_proj", |
| "note": "MLX-LM affine packing", |
| "success": true |
| } |
| ], |
| "schema_version": "axquant.quantizer-execution.v1" |
| } |
|
|
