{ "model_name": "deepseek-ai/deepseek-coder-6.7b-instruct", "output_format": "gptq", "target_bits": 4.0, "achieved_bits": 4.0, "majority_bits": 4, "symmetric": false, "sigma2": 0.001, "quant_granularity": "block", "block_cols": 64, "blocks": [ { "name": "model.embed_tokens.weight", "saliency": 0.0010000905022025108, "bits": 4, "params": 16515072 }, { "name": "model.layers.0.self_attn.q_proj.weight", "saliency": 0.0010000386973842978, "bits": 4, "params": 2097152 }, { "name": "model.layers.0.self_attn.k_proj.weight", "saliency": 0.0010000837501138449, "bits": 4, "params": 2097152 }, { "name": "model.layers.0.self_attn.v_proj.weight", "saliency": 0.001007600105367601, "bits": 4, "params": 2097152 }, { "name": "model.layers.0.self_attn.o_proj.weight", "saliency": 0.001002470264211297, "bits": 4, "params": 2097152 }, { "name": "model.layers.0.mlp.gate_proj.weight", "saliency": 0.001000305637717247, "bits": 4, "params": 5636096 }, { "name": "model.layers.0.mlp.up_proj.weight", "saliency": 0.0010003813076764345, "bits": 4, "params": 5636096 }, { "name": "model.layers.0.mlp.down_proj.weight", "saliency": 0.0010016071610152721, "bits": 4, "params": 5636096 }, { "name": "model.layers.0.input_layernorm.weight", "saliency": 0.0046534850262105465, "bits": 4, "params": 512 }, { "name": "model.layers.0.post_attention_layernorm.weight", "saliency": 0.001804268453270197, "bits": 4, "params": 512 }, { "name": "model.layers.1.self_attn.q_proj.weight", "saliency": 0.0010002334602177143, "bits": 4, "params": 2097152 }, { "name": "model.layers.1.self_attn.k_proj.weight", "saliency": 0.001000267919152975, "bits": 4, "params": 2097152 }, { "name": "model.layers.1.self_attn.v_proj.weight", "saliency": 0.0010432712733745575, "bits": 4, "params": 2097152 }, { "name": "model.layers.1.self_attn.o_proj.weight", "saliency": 0.0010016026208177209, "bits": 4, "params": 2097152 }, { "name": "model.layers.1.mlp.gate_proj.weight", "saliency": 0.0010001917835325003, "bits": 4, "params": 5636096 }, { "name": "model.layers.1.mlp.up_proj.weight", "saliency": 0.0010002562776207924, "bits": 4, "params": 5636096 }, { "name": "model.layers.1.mlp.down_proj.weight", "saliency": 0.0010021611815318465, "bits": 4, "params": 5636096 }, { "name": "model.layers.1.input_layernorm.weight", "saliency": 0.005320083349943161, "bits": 4, "params": 512 }, { "name": "model.layers.1.post_attention_layernorm.weight", "saliency": 0.003178255632519722, "bits": 4, "params": 512 }, { "name": "model.layers.2.self_attn.q_proj.weight", "saliency": 0.001000311691313982, "bits": 4, "params": 2097152 }, { "name": "model.layers.2.self_attn.k_proj.weight", "saliency": 0.0010002533672377467, "bits": 4, "params": 2097152 }, { "name": "model.layers.2.self_attn.v_proj.weight", "saliency": 0.001006864127703011, "bits": 4, "params": 2097152 }, { "name": "model.layers.2.self_attn.o_proj.weight", "saliency": 0.0010014860890805721, "bits": 4, "params": 2097152 }, { "name": "model.layers.2.mlp.gate_proj.weight", "saliency": 0.00100047851447016, "bits": 4, "params": 5636096 }, { "name": "model.layers.2.mlp.up_proj.weight", "saliency": 0.0010009685065597296, "bits": 4, "params": 5636096 }, { "name": "model.layers.2.mlp.down_proj.weight", "saliency": 0.001172472140751779, "bits": 4, "params": 5636096 }, { "name": "model.layers.2.input_layernorm.weight", "saliency": 0.0040357583202421665, "bits": 4, "params": 512 }, { "name": "model.layers.2.post_attention_layernorm.weight", "saliency": 0.012781282886862755, "bits": 4, "params": 512 }, { "name": "model.layers.3.self_attn.q_proj.weight", "saliency": 0.0010005959775298834, "bits": 4, "params": 2097152 }, { "name": "model.layers.3.self_attn.k_proj.weight", "saliency": 0.001000372227281332, "bits": 4, "params": 2097152 }, { "name": "model.layers.3.self_attn.v_proj.weight", "saliency": 0.0010035440791398287, "bits": 4, "params": 2097152 }, { "name": "model.layers.3.self_attn.o_proj.weight", "saliency": 0.0010013554710894823, "bits": 4, "params": 2097152 }, { "name": "model.layers.3.mlp.gate_proj.weight", "saliency": 0.0010003988863900304, "bits": 4, "params": 5636096 }, { "name": "model.layers.3.mlp.up_proj.weight", "saliency": 0.0010004491778090596, "bits": 4, "params": 5636096 }, { "name": "model.layers.3.mlp.down_proj.weight", "saliency": 0.0010005953954532743, "bits": 4, "params": 5636096 }, { "name": "model.layers.3.input_layernorm.weight", "saliency": 0.023367397487163544, "bits": 4, "params": 512 }, { "name": "model.layers.3.post_attention_layernorm.weight", "saliency": 0.005630176980048418, "bits": 4, "params": 512 }, { "name": "model.layers.4.self_attn.q_proj.weight", "saliency": 0.0010005788644775748, "bits": 4, "params": 2097152 }, { "name": "model.layers.4.self_attn.k_proj.weight", "saliency": 0.0010003214702010155, "bits": 4, "params": 2097152 }, { "name": "model.layers.4.self_attn.v_proj.weight", "saliency": 0.0010037433821707964, "bits": 4, "params": 2097152 }, { "name": "model.layers.4.self_attn.o_proj.weight", "saliency": 0.0010010522091761231, "bits": 4, "params": 2097152 }, { "name": "model.layers.4.mlp.gate_proj.weight", "saliency": 0.0010004715295508504, "bits": 4, "params": 5636096 }, { "name": "model.layers.4.mlp.up_proj.weight", "saliency": 0.0010004955111071467, "bits": 4, "params": 5636096 }, { "name": "model.layers.4.mlp.down_proj.weight", "saliency": 0.001000545802526176, "bits": 4, "params": 5636096 }, { "name": "model.layers.4.input_layernorm.weight", "saliency": 0.015614990144968033, "bits": 4, "params": 512 }, { "name": "model.layers.4.post_attention_layernorm.weight", "saliency": 0.004907149355858564, "bits": 4, "params": 512 }, { "name": "model.layers.5.self_attn.q_proj.weight", "saliency": 0.0010007246164605021, "bits": 4, "params": 2097152 }, { "name": "model.layers.5.self_attn.k_proj.weight", "saliency": 0.0010005810763686895, "bits": 4, "params": 2097152 }, { "name": "model.layers.5.self_attn.v_proj.weight", "saliency": 0.001004263642244041, "bits": 4, "params": 2097152 }, { "name": "model.layers.5.self_attn.o_proj.weight", "saliency": 0.0010006814263761044, "bits": 4, "params": 2097152 }, { "name": "model.layers.5.mlp.gate_proj.weight", "saliency": 0.0010004150681197643, "bits": 4, "params": 5636096 }, { "name": "model.layers.5.mlp.up_proj.weight", "saliency": 0.0010004935320466757, "bits": 4, "params": 5636096 }, { "name": "model.layers.5.mlp.down_proj.weight", "saliency": 0.0010005232179537416, "bits": 4, "params": 5636096 }, { "name": "model.layers.5.input_layernorm.weight", "saliency": 0.10245711356401443, "bits": 4, "params": 512 }, { "name": "model.layers.5.post_attention_layernorm.weight", "saliency": 0.0022739716805517673, "bits": 4, "params": 512 }, { "name": "model.layers.6.self_attn.q_proj.weight", "saliency": 0.0010007964447140694, "bits": 4, "params": 2097152 }, { "name": "model.layers.6.self_attn.k_proj.weight", "saliency": 0.0010004682699218392, "bits": 4, "params": 2097152 }, { "name": "model.layers.6.self_attn.v_proj.weight", "saliency": 0.0010037479223683476, "bits": 4, "params": 2097152 }, { "name": "model.layers.6.self_attn.o_proj.weight", "saliency": 0.001000777119770646, "bits": 4, "params": 2097152 }, { "name": "model.layers.6.mlp.gate_proj.weight", "saliency": 0.00100039504468441, "bits": 4, "params": 5636096 }, { "name": "model.layers.6.mlp.up_proj.weight", "saliency": 0.0010003948118537664, "bits": 4, "params": 5636096 }, { "name": "model.layers.6.mlp.down_proj.weight", "saliency": 0.0010003885254263878, "bits": 4, "params": 5636096 }, { "name": "model.layers.6.input_layernorm.weight", "saliency": 0.009596002288162708, "bits": 4, "params": 512 }, { "name": "model.layers.6.post_attention_layernorm.weight", "saliency": 0.0023789010010659695, "bits": 4, "params": 512 }, { "name": "model.layers.7.self_attn.q_proj.weight", "saliency": 0.001000644639134407, "bits": 4, "params": 2097152 }, { "name": "model.layers.7.self_attn.k_proj.weight", "saliency": 0.001000371528789401, "bits": 4, "params": 2097152 }, { "name": "model.layers.7.self_attn.v_proj.weight", "saliency": 0.0010040533961728215, "bits": 4, "params": 2097152 }, { "name": "model.layers.7.self_attn.o_proj.weight", "saliency": 0.0010006779339164495, "bits": 4, "params": 2097152 }, { "name": "model.layers.7.mlp.gate_proj.weight", "saliency": 0.001000349991954863, "bits": 4, "params": 5636096 }, { "name": "model.layers.7.mlp.up_proj.weight", "saliency": 0.0010004147188737988, "bits": 4, "params": 5636096 }, { "name": "model.layers.7.mlp.down_proj.weight", "saliency": 0.0010003725765272975, "bits": 4, "params": 5636096 }, { "name": "model.layers.7.input_layernorm.weight", "saliency": 0.006539833731949329, "bits": 4, "params": 512 }, { "name": "model.layers.7.post_attention_layernorm.weight", "saliency": 0.0022120883222669363, "bits": 4, "params": 512 }, { "name": "model.layers.8.self_attn.q_proj.weight", "saliency": 0.0010009788675233722, "bits": 4, "params": 2097152 }, { "name": "model.layers.8.self_attn.k_proj.weight", "saliency": 0.0010005614021793008, "bits": 4, "params": 2097152 }, { "name": "model.layers.8.self_attn.v_proj.weight", "saliency": 0.0010037095053121448, "bits": 4, "params": 2097152 }, { "name": "model.layers.8.self_attn.o_proj.weight", "saliency": 0.0010006353259086609, "bits": 4, "params": 2097152 }, { "name": "model.layers.8.mlp.gate_proj.weight", "saliency": 0.0010003331117331982, "bits": 4, "params": 5636096 }, { "name": "model.layers.8.mlp.up_proj.weight", "saliency": 0.0010003886418417096, "bits": 4, "params": 5636096 }, { "name": "model.layers.8.mlp.down_proj.weight", "saliency": 0.0010003066854551435, "bits": 4, "params": 5636096 }, { "name": "model.layers.8.input_layernorm.weight", "saliency": 0.002413144800812006, "bits": 4, "params": 512 }, { "name": "model.layers.8.post_attention_layernorm.weight", "saliency": 0.0019250639015808702, "bits": 4, "params": 512 }, { "name": "model.layers.9.self_attn.q_proj.weight", "saliency": 0.0010010547703132033, "bits": 4, "params": 2097152 }, { "name": "model.layers.9.self_attn.k_proj.weight", "saliency": 0.0010003905044868588, "bits": 4, "params": 2097152 }, { "name": "model.layers.9.self_attn.v_proj.weight", "saliency": 0.0010036980966106057, "bits": 4, "params": 2097152 }, { "name": "model.layers.9.self_attn.o_proj.weight", "saliency": 0.0010007781675085425, "bits": 4, "params": 2097152 }, { "name": "model.layers.9.mlp.gate_proj.weight", "saliency": 0.0010002597700804472, "bits": 4, "params": 5636096 }, { "name": "model.layers.9.mlp.up_proj.weight", "saliency": 0.0010002567432820797, "bits": 4, "params": 5636096 }, { "name": "model.layers.9.mlp.down_proj.weight", "saliency": 0.0010001822374761105, "bits": 4, "params": 5636096 }, { "name": "model.layers.9.input_layernorm.weight", "saliency": 0.008356169797480106, "bits": 4, "params": 512 }, { "name": "model.layers.9.post_attention_layernorm.weight", "saliency": 0.002310697454959154, "bits": 4, "params": 512 }, { "name": "model.layers.10.self_attn.q_proj.weight", "saliency": 0.0010016688611358404, "bits": 4, "params": 2097152 }, { "name": "model.layers.10.self_attn.k_proj.weight", "saliency": 0.0010009293910115957, "bits": 4, "params": 2097152 }, { "name": "model.layers.10.self_attn.v_proj.weight", "saliency": 0.001001864904537797, "bits": 4, "params": 2097152 }, { "name": "model.layers.10.self_attn.o_proj.weight", "saliency": 0.001000389689579606, "bits": 4, "params": 2097152 }, { "name": "model.layers.10.mlp.gate_proj.weight", "saliency": 0.0010001379996538162, "bits": 4, "params": 5636096 }, { "name": "model.layers.10.mlp.up_proj.weight", "saliency": 0.0010001371847465634, "bits": 4, "params": 5636096 }, { "name": "model.layers.10.mlp.down_proj.weight", "saliency": 0.0010000872425734997, "bits": 4, "params": 5636096 }, { "name": "model.layers.10.input_layernorm.weight", "saliency": 0.0021631959825754166, "bits": 4, "params": 512 }, { "name": "model.layers.10.post_attention_layernorm.weight", "saliency": 0.0022094445303082466, "bits": 4, "params": 512 }, { "name": "model.layers.11.self_attn.q_proj.weight", "saliency": 0.0010002595372498035, "bits": 4, "params": 2097152 }, { "name": "model.layers.11.self_attn.k_proj.weight", "saliency": 0.0010002534836530685, "bits": 4, "params": 2097152 }, { "name": "model.layers.11.self_attn.v_proj.weight", "saliency": 0.0010006153024733067, "bits": 4, "params": 2097152 }, { "name": "model.layers.11.self_attn.o_proj.weight", "saliency": 0.0010000724578276277, "bits": 4, "params": 2097152 }, { "name": "model.layers.11.mlp.gate_proj.weight", "saliency": 0.0010000963229686022, "bits": 4, "params": 5636096 }, { "name": "model.layers.11.mlp.up_proj.weight", "saliency": 0.001000179792754352, "bits": 4, "params": 5636096 }, { "name": "model.layers.11.mlp.down_proj.weight", "saliency": 0.0010093104792758822, "bits": 4, "params": 5636096 }, { "name": "model.layers.11.input_layernorm.weight", "saliency": 0.0031055486761033535, "bits": 4, "params": 512 }, { "name": "model.layers.11.post_attention_layernorm.weight", "saliency": 0.0015748429577797651, "bits": 4, "params": 512 }, { "name": "model.layers.12.self_attn.q_proj.weight", "saliency": 0.0010000135516747832, "bits": 4, "params": 2097152 }, { "name": "model.layers.12.self_attn.k_proj.weight", "saliency": 0.0010000087786465883, "bits": 4, "params": 2097152 }, { "name": "model.layers.12.self_attn.v_proj.weight", "saliency": 0.0010004483629018068, "bits": 4, "params": 2097152 }, { "name": "model.layers.12.self_attn.o_proj.weight", "saliency": 0.0010000703623518348, "bits": 4, "params": 2097152 }, { "name": "model.layers.12.mlp.gate_proj.weight", "saliency": 0.0010000292677432299, "bits": 4, "params": 5636096 }, { "name": "model.layers.12.mlp.up_proj.weight", "saliency": 0.0010000306647270918, "bits": 4, "params": 5636096 }, { "name": "model.layers.12.mlp.down_proj.weight", "saliency": 0.0010000800248235464, "bits": 4, "params": 5636096 }, { "name": "model.layers.12.input_layernorm.weight", "saliency": 0.0010588932782411575, "bits": 4, "params": 512 }, { "name": "model.layers.12.post_attention_layernorm.weight", "saliency": 0.0034138173796236515, "bits": 4, "params": 512 }, { "name": "model.layers.13.self_attn.q_proj.weight", "saliency": 0.0010000125039368868, "bits": 4, "params": 2097152 }, { "name": "model.layers.13.self_attn.k_proj.weight", "saliency": 0.0010000092443078756, "bits": 4, "params": 2097152 }, { "name": "model.layers.13.self_attn.v_proj.weight", "saliency": 0.0010004998184740543, "bits": 4, "params": 2097152 }, { "name": "model.layers.13.self_attn.o_proj.weight", "saliency": 0.0010000475449487567, "bits": 4, "params": 2097152 }, { "name": "model.layers.13.mlp.gate_proj.weight", "saliency": 0.0010000214679166675, "bits": 4, "params": 5636096 }, { "name": "model.layers.13.mlp.up_proj.weight", "saliency": 0.0010000235633924603, "bits": 4, "params": 5636096 }, { "name": "model.layers.13.mlp.down_proj.weight", "saliency": 0.0010000393958762288, "bits": 4, "params": 5636096 }, { "name": "model.layers.13.input_layernorm.weight", "saliency": 0.0010296208783984184, "bits": 4, "params": 512 }, { "name": "model.layers.13.post_attention_layernorm.weight", "saliency": 0.001062327646650374, "bits": 4, "params": 512 }, { "name": "model.layers.14.self_attn.q_proj.weight", "saliency": 0.001000010408461094, "bits": 4, "params": 2097152 }, { "name": "model.layers.14.self_attn.k_proj.weight", "saliency": 0.0010000059846788645, "bits": 4, "params": 2097152 }, { "name": "model.layers.14.self_attn.v_proj.weight", "saliency": 0.0010004452196881175, "bits": 4, "params": 2097152 }, { "name": "model.layers.14.self_attn.o_proj.weight", "saliency": 0.0010000512702390552, "bits": 4, "params": 2097152 }, { "name": "model.layers.14.mlp.gate_proj.weight", "saliency": 0.001000022515654564, "bits": 4, "params": 5636096 }, { "name": "model.layers.14.mlp.up_proj.weight", "saliency": 0.0010000228649005294, "bits": 4, "params": 5636096 }, { "name": "model.layers.14.mlp.down_proj.weight", "saliency": 0.0010000356705859303, "bits": 4, "params": 5636096 }, { "name": "model.layers.14.input_layernorm.weight", "saliency": 0.0010283105075359344, "bits": 4, "params": 512 }, { "name": "model.layers.14.post_attention_layernorm.weight", "saliency": 0.001033260952681303, "bits": 4, "params": 512 }, { "name": "model.layers.15.self_attn.q_proj.weight", "saliency": 0.0010000286856666207, "bits": 4, "params": 2097152 }, { "name": "model.layers.15.self_attn.k_proj.weight", "saliency": 0.0010000171605497599, "bits": 4, "params": 2097152 }, { "name": "model.layers.15.self_attn.v_proj.weight", "saliency": 0.0010006362572312355, "bits": 4, "params": 2097152 }, { "name": "model.layers.15.self_attn.o_proj.weight", "saliency": 0.001000061514787376, "bits": 4, "params": 2097152 }, { "name": "model.layers.15.mlp.gate_proj.weight", "saliency": 0.0010000234469771385, "bits": 4, "params": 5636096 }, { "name": "model.layers.15.mlp.up_proj.weight", "saliency": 0.0010000241454690695, "bits": 4, "params": 5636096 }, { "name": "model.layers.15.mlp.down_proj.weight", "saliency": 0.0010000384645536542, "bits": 4, "params": 5636096 }, { "name": "model.layers.15.input_layernorm.weight", "saliency": 0.0010344466427341104, "bits": 4, "params": 512 }, { "name": "model.layers.15.post_attention_layernorm.weight", "saliency": 0.001027294434607029, "bits": 4, "params": 512 }, { "name": "model.layers.16.self_attn.q_proj.weight", "saliency": 0.0010000247275456786, "bits": 4, "params": 2097152 }, { "name": "model.layers.16.self_attn.k_proj.weight", "saliency": 0.0010000147158280015, "bits": 4, "params": 2097152 }, { "name": "model.layers.16.self_attn.v_proj.weight", "saliency": 0.0010004955111071467, "bits": 4, "params": 2097152 }, { "name": "model.layers.16.self_attn.o_proj.weight", "saliency": 0.0010000610491260886, "bits": 4, "params": 2097152 }, { "name": "model.layers.16.mlp.gate_proj.weight", "saliency": 0.001000025193206966, "bits": 4, "params": 5636096 }, { "name": "model.layers.16.mlp.up_proj.weight", "saliency": 0.0010000261245295405, "bits": 4, "params": 5636096 }, { "name": "model.layers.16.mlp.down_proj.weight", "saliency": 0.001000046031549573, "bits": 4, "params": 5636096 }, { "name": "model.layers.16.input_layernorm.weight", "saliency": 0.0010417994344606996, "bits": 4, "params": 512 }, { "name": "model.layers.16.post_attention_layernorm.weight", "saliency": 0.0010264910524711013, "bits": 4, "params": 512 }, { "name": "model.layers.17.self_attn.q_proj.weight", "saliency": 0.0010000127367675304, "bits": 4, "params": 2097152 }, { "name": "model.layers.17.self_attn.k_proj.weight", "saliency": 0.0010000099427998066, "bits": 4, "params": 2097152 }, { "name": "model.layers.17.self_attn.v_proj.weight", "saliency": 0.001000493299216032, "bits": 4, "params": 2097152 }, { "name": "model.layers.17.self_attn.o_proj.weight", "saliency": 0.001000037183985114, "bits": 4, "params": 2097152 }, { "name": "model.layers.17.mlp.gate_proj.weight", "saliency": 0.001000039279460907, "bits": 4, "params": 5636096 }, { "name": "model.layers.17.mlp.up_proj.weight", "saliency": 0.0010000390466302633, "bits": 4, "params": 5636096 }, { "name": "model.layers.17.mlp.down_proj.weight", "saliency": 0.0010001558111980557, "bits": 4, "params": 5636096 }, { "name": "model.layers.17.input_layernorm.weight", "saliency": 0.0010452896822243929, "bits": 4, "params": 512 }, { "name": "model.layers.17.post_attention_layernorm.weight", "saliency": 0.005208131857216358, "bits": 4, "params": 512 }, { "name": "model.layers.18.self_attn.q_proj.weight", "saliency": 0.001000011689029634, "bits": 4, "params": 2097152 }, { "name": "model.layers.18.self_attn.k_proj.weight", "saliency": 0.001000009011477232, "bits": 4, "params": 2097152 }, { "name": "model.layers.18.self_attn.v_proj.weight", "saliency": 0.0010004055220633745, "bits": 4, "params": 2097152 }, { "name": "model.layers.18.self_attn.o_proj.weight", "saliency": 0.0010000321781262755, "bits": 4, "params": 2097152 }, { "name": "model.layers.18.mlp.gate_proj.weight", "saliency": 0.0010000306647270918, "bits": 4, "params": 5636096 }, { "name": "model.layers.18.mlp.up_proj.weight", "saliency": 0.0010000303154811263, "bits": 4, "params": 5636096 }, { "name": "model.layers.18.mlp.down_proj.weight", "saliency": 0.0010000438196584582, "bits": 4, "params": 5636096 }, { "name": "model.layers.18.input_layernorm.weight", "saliency": 0.0010255861561745405, "bits": 4, "params": 512 }, { "name": "model.layers.18.post_attention_layernorm.weight", "saliency": 0.0010326086776331067, "bits": 4, "params": 512 }, { "name": "model.layers.19.self_attn.q_proj.weight", "saliency": 0.0010000168113037944, "bits": 4, "params": 2097152 }, { "name": "model.layers.19.self_attn.k_proj.weight", "saliency": 0.0010000108741223812, "bits": 4, "params": 2097152 }, { "name": "model.layers.19.self_attn.v_proj.weight", "saliency": 0.0010004359064623713, "bits": 4, "params": 2097152 }, { "name": "model.layers.19.self_attn.o_proj.weight", "saliency": 0.001000021817162633, "bits": 4, "params": 2097152 }, { "name": "model.layers.19.mlp.gate_proj.weight", "saliency": 0.0010000340407714248, "bits": 4, "params": 5636096 }, { "name": "model.layers.19.mlp.up_proj.weight", "saliency": 0.0010000314796343446, "bits": 4, "params": 5636096 }, { "name": "model.layers.19.mlp.down_proj.weight", "saliency": 0.0010000403271988034, "bits": 4, "params": 5636096 }, { "name": "model.layers.19.input_layernorm.weight", "saliency": 0.0010206678416579962, "bits": 4, "params": 512 }, { "name": "model.layers.19.post_attention_layernorm.weight", "saliency": 0.0011190972290933132, "bits": 4, "params": 512 }, { "name": "model.layers.20.self_attn.q_proj.weight", "saliency": 0.001000009011477232, "bits": 4, "params": 2097152 }, { "name": "model.layers.20.self_attn.k_proj.weight", "saliency": 0.0010000114561989903, "bits": 4, "params": 2097152 }, { "name": "model.layers.20.self_attn.v_proj.weight", "saliency": 0.0010004123905673623, "bits": 4, "params": 2097152 }, { "name": "model.layers.20.self_attn.o_proj.weight", "saliency": 0.0010000187903642654, "bits": 4, "params": 2097152 }, { "name": "model.layers.20.mlp.gate_proj.weight", "saliency": 0.0010000343900173903, "bits": 4, "params": 5636096 }, { "name": "model.layers.20.mlp.up_proj.weight", "saliency": 0.001000034506432712, "bits": 4, "params": 5636096 }, { "name": "model.layers.20.mlp.down_proj.weight", "saliency": 0.0010000390466302633, "bits": 4, "params": 5636096 }, { "name": "model.layers.20.input_layernorm.weight", "saliency": 0.001019699266180396, "bits": 4, "params": 512 }, { "name": "model.layers.20.post_attention_layernorm.weight", "saliency": 0.0010321969166398048, "bits": 4, "params": 512 }, { "name": "model.layers.21.self_attn.q_proj.weight", "saliency": 0.0010000363690778613, "bits": 4, "params": 2097152 }, { "name": "model.layers.21.self_attn.k_proj.weight", "saliency": 0.0010000194888561964, "bits": 4, "params": 2097152 }, { "name": "model.layers.21.self_attn.v_proj.weight", "saliency": 0.0010003469651564956, "bits": 4, "params": 2097152 }, { "name": "model.layers.21.self_attn.o_proj.weight", "saliency": 0.0010000222828239202, "bits": 4, "params": 2097152 }, { "name": "model.layers.21.mlp.gate_proj.weight", "saliency": 0.0010000377660617232, "bits": 4, "params": 5636096 }, { "name": "model.layers.21.mlp.up_proj.weight", "saliency": 0.0010000370675697923, "bits": 4, "params": 5636096 }, { "name": "model.layers.21.mlp.down_proj.weight", "saliency": 0.001000034622848034, "bits": 4, "params": 5636096 }, { "name": "model.layers.21.input_layernorm.weight", "saliency": 0.001093475497327745, "bits": 4, "params": 512 }, { "name": "model.layers.21.post_attention_layernorm.weight", "saliency": 0.0010208565508946776, "bits": 4, "params": 512 }, { "name": "model.layers.22.self_attn.q_proj.weight", "saliency": 0.0010000118054449558, "bits": 4, "params": 2097152 }, { "name": "model.layers.22.self_attn.k_proj.weight", "saliency": 0.0010000088950619102, "bits": 4, "params": 2097152 }, { "name": "model.layers.22.self_attn.v_proj.weight", "saliency": 0.001000461052171886, "bits": 4, "params": 2097152 }, { "name": "model.layers.22.self_attn.o_proj.weight", "saliency": 0.001000037183985114, "bits": 4, "params": 2097152 }, { "name": "model.layers.22.mlp.gate_proj.weight", "saliency": 0.001000041258521378, "bits": 4, "params": 5636096 }, { "name": "model.layers.22.mlp.up_proj.weight", "saliency": 0.0010000396287068725, "bits": 4, "params": 5636096 }, { "name": "model.layers.22.mlp.down_proj.weight", "saliency": 0.0010000320617109537, "bits": 4, "params": 5636096 }, { "name": "model.layers.22.input_layernorm.weight", "saliency": 0.0010445659281685948, "bits": 4, "params": 512 }, { "name": "model.layers.22.post_attention_layernorm.weight", "saliency": 0.0010544650722295046, "bits": 4, "params": 512 }, { "name": "model.layers.23.self_attn.q_proj.weight", "saliency": 0.0010000286856666207, "bits": 4, "params": 2097152 }, { "name": "model.layers.23.self_attn.k_proj.weight", "saliency": 0.001000023097731173, "bits": 4, "params": 2097152 }, { "name": "model.layers.23.self_attn.v_proj.weight", "saliency": 0.0010005098301917315, "bits": 4, "params": 2097152 }, { "name": "model.layers.23.self_attn.o_proj.weight", "saliency": 0.001000028452835977, "bits": 4, "params": 2097152 }, { "name": "model.layers.23.mlp.gate_proj.weight", "saliency": 0.0010000469628721476, "bits": 4, "params": 5636096 }, { "name": "model.layers.23.mlp.up_proj.weight", "saliency": 0.0010000428883358836, "bits": 4, "params": 5636096 }, { "name": "model.layers.23.mlp.down_proj.weight", "saliency": 0.001000033225864172, "bits": 4, "params": 5636096 }, { "name": "model.layers.23.input_layernorm.weight", "saliency": 0.0010116384364664555, "bits": 4, "params": 512 }, { "name": "model.layers.23.post_attention_layernorm.weight", "saliency": 0.0013022581115365028, "bits": 4, "params": 512 }, { "name": "model.layers.24.self_attn.q_proj.weight", "saliency": 0.0010000080801546574, "bits": 4, "params": 2097152 }, { "name": "model.layers.24.self_attn.k_proj.weight", "saliency": 0.0010000065667554736, "bits": 4, "params": 2097152 }, { "name": "model.layers.24.self_attn.v_proj.weight", "saliency": 0.0010004322975873947, "bits": 4, "params": 2097152 }, { "name": "model.layers.24.self_attn.o_proj.weight", "saliency": 0.0010000185575336218, "bits": 4, "params": 2097152 }, { "name": "model.layers.24.mlp.gate_proj.weight", "saliency": 0.001000039977952838, "bits": 4, "params": 5636096 }, { "name": "model.layers.24.mlp.up_proj.weight", "saliency": 0.001000039977952838, "bits": 4, "params": 5636096 }, { "name": "model.layers.24.mlp.down_proj.weight", "saliency": 0.0010000213515013456, "bits": 4, "params": 5636096 }, { "name": "model.layers.24.input_layernorm.weight", "saliency": 0.0010260919807478786, "bits": 4, "params": 512 }, { "name": "model.layers.24.post_attention_layernorm.weight", "saliency": 0.0010397722944617271, "bits": 4, "params": 512 }, { "name": "model.layers.25.self_attn.q_proj.weight", "saliency": 0.0010000114561989903, "bits": 4, "params": 2097152 }, { "name": "model.layers.25.self_attn.k_proj.weight", "saliency": 0.00100000761449337, "bits": 4, "params": 2097152 }, { "name": "model.layers.25.self_attn.v_proj.weight", "saliency": 0.0010003685019910336, "bits": 4, "params": 2097152 }, { "name": "model.layers.25.self_attn.o_proj.weight", "saliency": 0.001000016462057829, "bits": 4, "params": 2097152 }, { "name": "model.layers.25.mlp.gate_proj.weight", "saliency": 0.0010000384645536542, "bits": 4, "params": 5636096 }, { "name": "model.layers.25.mlp.up_proj.weight", "saliency": 0.0010000288020819426, "bits": 4, "params": 5636096 }, { "name": "model.layers.25.mlp.down_proj.weight", "saliency": 0.0010000207694247365, "bits": 4, "params": 5636096 }, { "name": "model.layers.25.input_layernorm.weight", "saliency": 0.0010052515426650643, "bits": 4, "params": 512 }, { "name": "model.layers.25.post_attention_layernorm.weight", "saliency": 0.0010316011030226946, "bits": 4, "params": 512 }, { "name": "model.layers.26.self_attn.q_proj.weight", "saliency": 0.0010000132024288177, "bits": 4, "params": 2097152 }, { "name": "model.layers.26.self_attn.k_proj.weight", "saliency": 0.0010000128531828523, "bits": 4, "params": 2097152 }, { "name": "model.layers.26.self_attn.v_proj.weight", "saliency": 0.001000437419861555, "bits": 4, "params": 2097152 }, { "name": "model.layers.26.self_attn.o_proj.weight", "saliency": 0.001000022515654564, "bits": 4, "params": 2097152 }, { "name": "model.layers.26.mlp.gate_proj.weight", "saliency": 0.001000025076791644, "bits": 4, "params": 5636096 }, { "name": "model.layers.26.mlp.up_proj.weight", "saliency": 0.0010000234469771385, "bits": 4, "params": 5636096 }, { "name": "model.layers.26.mlp.down_proj.weight", "saliency": 0.0010000134352594614, "bits": 4, "params": 5636096 }, { "name": "model.layers.26.input_layernorm.weight", "saliency": 0.0010503950761631131, "bits": 4, "params": 512 }, { "name": "model.layers.26.post_attention_layernorm.weight", "saliency": 0.001027348916977644, "bits": 4, "params": 512 }, { "name": "model.layers.27.self_attn.q_proj.weight", "saliency": 0.0010000084294006228, "bits": 4, "params": 2097152 }, { "name": "model.layers.27.self_attn.k_proj.weight", "saliency": 0.001000011689029634, "bits": 4, "params": 2097152 }, { "name": "model.layers.27.self_attn.v_proj.weight", "saliency": 0.0010005002841353416, "bits": 4, "params": 2097152 }, { "name": "model.layers.27.self_attn.o_proj.weight", "saliency": 0.0010000168113037944, "bits": 4, "params": 2097152 }, { "name": "model.layers.27.mlp.gate_proj.weight", "saliency": 0.0010000228649005294, "bits": 4, "params": 5636096 }, { "name": "model.layers.27.mlp.up_proj.weight", "saliency": 0.0010000180918723345, "bits": 4, "params": 5636096 }, { "name": "model.layers.27.mlp.down_proj.weight", "saliency": 0.0010000099427998066, "bits": 4, "params": 5636096 }, { "name": "model.layers.27.input_layernorm.weight", "saliency": 0.0010269742924720049, "bits": 4, "params": 512 }, { "name": "model.layers.27.post_attention_layernorm.weight", "saliency": 0.0010558314388617873, "bits": 4, "params": 512 }, { "name": "model.layers.28.self_attn.q_proj.weight", "saliency": 0.0010000107577070594, "bits": 4, "params": 2097152 }, { "name": "model.layers.28.self_attn.k_proj.weight", "saliency": 0.0010000121546909213, "bits": 4, "params": 2097152 }, { "name": "model.layers.28.self_attn.v_proj.weight", "saliency": 0.0010005213553085923, "bits": 4, "params": 2097152 }, { "name": "model.layers.28.self_attn.o_proj.weight", "saliency": 0.0010000248439610004, "bits": 4, "params": 2097152 }, { "name": "model.layers.28.mlp.gate_proj.weight", "saliency": 0.001000025193206966, "bits": 4, "params": 5636096 }, { "name": "model.layers.28.mlp.up_proj.weight", "saliency": 0.0010001881746575236, "bits": 4, "params": 5636096 }, { "name": "model.layers.28.mlp.down_proj.weight", "saliency": 0.0010000162292271852, "bits": 4, "params": 5636096 }, { "name": "model.layers.28.input_layernorm.weight", "saliency": 0.0010444106301292777, "bits": 4, "params": 512 }, { "name": "model.layers.28.post_attention_layernorm.weight", "saliency": 0.006702560931444168, "bits": 4, "params": 512 }, { "name": "model.layers.29.self_attn.q_proj.weight", "saliency": 0.0010000147158280015, "bits": 4, "params": 2097152 }, { "name": "model.layers.29.self_attn.k_proj.weight", "saliency": 0.001000019139610231, "bits": 4, "params": 2097152 }, { "name": "model.layers.29.self_attn.v_proj.weight", "saliency": 0.0010008614044636488, "bits": 4, "params": 2097152 }, { "name": "model.layers.29.self_attn.o_proj.weight", "saliency": 0.0010000538313761353, "bits": 4, "params": 2097152 }, { "name": "model.layers.29.mlp.gate_proj.weight", "saliency": 0.0010005640797317028, "bits": 4, "params": 5636096 }, { "name": "model.layers.29.mlp.up_proj.weight", "saliency": 0.001000583521090448, "bits": 4, "params": 5636096 }, { "name": "model.layers.29.mlp.down_proj.weight", "saliency": 0.0010354513069614768, "bits": 4, "params": 5636096 }, { "name": "model.layers.29.input_layernorm.weight", "saliency": 0.001268706633709371, "bits": 4, "params": 512 }, { "name": "model.layers.29.post_attention_layernorm.weight", "saliency": 0.013578707352280617, "bits": 4, "params": 512 }, { "name": "model.layers.30.self_attn.q_proj.weight", "saliency": 0.00100000633392483, "bits": 4, "params": 2097152 }, { "name": "model.layers.30.self_attn.k_proj.weight", "saliency": 0.0010000118054449558, "bits": 4, "params": 2097152 }, { "name": "model.layers.30.self_attn.v_proj.weight", "saliency": 0.001000517513602972, "bits": 4, "params": 2097152 }, { "name": "model.layers.30.self_attn.o_proj.weight", "saliency": 0.0010001164628192782, "bits": 4, "params": 2097152 }, { "name": "model.layers.30.mlp.gate_proj.weight", "saliency": 0.0010001598857343197, "bits": 4, "params": 5636096 }, { "name": "model.layers.30.mlp.up_proj.weight", "saliency": 0.0010001171613112092, "bits": 4, "params": 5636096 }, { "name": "model.layers.30.mlp.down_proj.weight", "saliency": 0.001000144868157804, "bits": 4, "params": 5636096 }, { "name": "model.layers.30.input_layernorm.weight", "saliency": 0.0010595009662210941, "bits": 4, "params": 512 }, { "name": "model.layers.30.post_attention_layernorm.weight", "saliency": 0.00192739418707788, "bits": 4, "params": 512 }, { "name": "model.layers.31.self_attn.q_proj.weight", "saliency": 0.001000012969598174, "bits": 4, "params": 2097152 }, { "name": "model.layers.31.self_attn.k_proj.weight", "saliency": 0.0010000213515013456, "bits": 4, "params": 2097152 }, { "name": "model.layers.31.self_attn.v_proj.weight", "saliency": 0.0010003054048866034, "bits": 4, "params": 2097152 }, { "name": "model.layers.31.self_attn.o_proj.weight", "saliency": 0.0010001069167628884, "bits": 4, "params": 2097152 }, { "name": "model.layers.31.mlp.gate_proj.weight", "saliency": 0.0010001759510487318, "bits": 4, "params": 5636096 }, { "name": "model.layers.31.mlp.up_proj.weight", "saliency": 0.001000091666355729, "bits": 4, "params": 5636096 }, { "name": "model.layers.31.mlp.down_proj.weight", "saliency": 0.0010004746727645397, "bits": 4, "params": 5636096 }, { "name": "model.layers.31.input_layernorm.weight", "saliency": 0.001331512350589037, "bits": 4, "params": 512 }, { "name": "model.layers.31.post_attention_layernorm.weight", "saliency": 0.0020043200347572565, "bits": 4, "params": 512 }, { "name": "model.norm.weight", "saliency": 0.0015174412401393056, "bits": 4, "params": 512 }, { "name": "lm_head.weight", "saliency": 0.0010000175097957253, "bits": 4, "params": 16515072 } ] }