Hugging Face's logo

GreenBitAI
/
01-Yi-6B-Chat-layer-mix-bpw-2.2

Text Generation
Transformers
Safetensors
llama
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
NicoNico6 commited on
Commit
ff6c335
·
1 Parent(s): 04f290b

update

Browse files
Files changed (2) hide show
  1. model.safetensors +2 -2
  2. quant_strategy.json +54 -54
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:91a82dd7b9d7bdc723872d4e740a55f6ddf332167e78b58a65c4fafadf40c039
3
- size 2868360417
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad86ccda8515329f8899d598073269c8a4930ba5a8f7090fa4b591765c4b9ad7
3
+ size 2867900760
quant_strategy.json CHANGED
@@ -89,8 +89,8 @@
89
  }
90
  },
91
  "model.layers.1": {
92
- "accuracy": 0.9216856956481934,
93
- "total_bits": 486917664,
94
  "q_proj": {
95
  "group_size": {
96
  "2": 64
@@ -129,10 +129,10 @@
129
  },
130
  "o_proj": {
131
  "group_size": {
132
- "4": 128
133
  },
134
  "bits": [
135
- 4
136
  ],
137
  "bits_prop": [
138
  1
@@ -165,10 +165,10 @@
165
  },
166
  "down_proj": {
167
  "group_size": {
168
- "4": 128
169
  },
170
  "bits": [
171
- 4
172
  ],
173
  "bits_prop": [
174
  1
@@ -265,8 +265,8 @@
265
  }
266
  },
267
  "model.layers.3": {
268
- "accuracy": 0.9084997177124023,
269
- "total_bits": 449761824,
270
  "q_proj": {
271
  "group_size": {
272
  "2": 64
@@ -341,10 +341,10 @@
341
  },
342
  "down_proj": {
343
  "group_size": {
344
- "4": 128
345
  },
346
  "bits": [
347
- 4
348
  ],
349
  "bits_prop": [
350
  1
@@ -705,8 +705,8 @@
705
  }
706
  },
707
  "model.layers.8": {
708
- "accuracy": 0.9228010177612305,
709
- "total_bits": 449761824,
710
  "q_proj": {
711
  "group_size": {
712
  "2": 64
@@ -781,10 +781,10 @@
781
  },
782
  "down_proj": {
783
  "group_size": {
784
- "4": 128
785
  },
786
  "bits": [
787
- 4
788
  ],
789
  "bits_prop": [
790
  1
@@ -793,8 +793,8 @@
793
  }
794
  },
795
  "model.layers.9": {
796
- "accuracy": 0.9557886123657227,
797
- "total_bits": 360997440,
798
  "q_proj": {
799
  "group_size": {
800
  "2": 64
@@ -809,10 +809,10 @@
809
  },
810
  "k_proj": {
811
  "group_size": {
812
- "2": 64
813
  },
814
  "bits": [
815
- 2
816
  ],
817
  "bits_prop": [
818
  1
@@ -1057,8 +1057,8 @@
1057
  }
1058
  },
1059
  "model.layers.12": {
1060
- "accuracy": 0.9378215074539185,
1061
- "total_bits": 360997440,
1062
  "q_proj": {
1063
  "group_size": {
1064
  "2": 64
@@ -1073,10 +1073,10 @@
1073
  },
1074
  "k_proj": {
1075
  "group_size": {
1076
- "2": 64
1077
  },
1078
  "bits": [
1079
- 2
1080
  ],
1081
  "bits_prop": [
1082
  1
@@ -1233,8 +1233,8 @@
1233
  }
1234
  },
1235
  "model.layers.14": {
1236
- "accuracy": 0.9317333698272705,
1237
- "total_bits": 360997440,
1238
  "q_proj": {
1239
  "group_size": {
1240
  "2": 64
@@ -1249,10 +1249,10 @@
1249
  },
1250
  "k_proj": {
1251
  "group_size": {
1252
- "2": 64
1253
  },
1254
  "bits": [
1255
- 2
1256
  ],
1257
  "bits_prop": [
1258
  1
@@ -1761,8 +1761,8 @@
1761
  }
1762
  },
1763
  "model.layers.20": {
1764
- "accuracy": 0.9032342433929443,
1765
- "total_bits": 360997440,
1766
  "q_proj": {
1767
  "group_size": {
1768
  "2": 64
@@ -1777,10 +1777,10 @@
1777
  },
1778
  "k_proj": {
1779
  "group_size": {
1780
- "2": 64
1781
  },
1782
  "bits": [
1783
- 2
1784
  ],
1785
  "bits_prop": [
1786
  1
@@ -1849,8 +1849,8 @@
1849
  }
1850
  },
1851
  "model.layers.21": {
1852
- "accuracy": 0.9000318050384521,
1853
- "total_bits": 360997440,
1854
  "q_proj": {
1855
  "group_size": {
1856
  "2": 64
@@ -1865,10 +1865,10 @@
1865
  },
1866
  "k_proj": {
1867
  "group_size": {
1868
- "2": 64
1869
  },
1870
  "bits": [
1871
- 2
1872
  ],
1873
  "bits_prop": [
1874
  1
@@ -2289,8 +2289,8 @@
2289
  }
2290
  },
2291
  "model.layers.26": {
2292
- "accuracy": 0.9015021324157715,
2293
- "total_bits": 360997440,
2294
  "q_proj": {
2295
  "group_size": {
2296
  "2": 64
@@ -2305,10 +2305,10 @@
2305
  },
2306
  "k_proj": {
2307
  "group_size": {
2308
- "2": 64
2309
  },
2310
  "bits": [
2311
- 2
2312
  ],
2313
  "bits_prop": [
2314
  1
@@ -2465,8 +2465,8 @@
2465
  }
2466
  },
2467
  "model.layers.28": {
2468
- "accuracy": 0.9067087173461914,
2469
- "total_bits": 360997440,
2470
  "q_proj": {
2471
  "group_size": {
2472
  "2": 64
@@ -2541,10 +2541,10 @@
2541
  },
2542
  "down_proj": {
2543
  "group_size": {
2544
- "2": 64
2545
  },
2546
  "bits": [
2547
- 2
2548
  ],
2549
  "bits_prop": [
2550
  1
@@ -2553,8 +2553,8 @@
2553
  }
2554
  },
2555
  "model.layers.29": {
2556
- "accuracy": 0.9141254425048828,
2557
- "total_bits": 360997440,
2558
  "q_proj": {
2559
  "group_size": {
2560
  "2": 64
@@ -2629,10 +2629,10 @@
2629
  },
2630
  "down_proj": {
2631
  "group_size": {
2632
- "2": 64
2633
  },
2634
  "bits": [
2635
- 2
2636
  ],
2637
  "bits_prop": [
2638
  1
@@ -2641,8 +2641,8 @@
2641
  }
2642
  },
2643
  "model.layers.30": {
2644
- "accuracy": 0.9168109893798828,
2645
- "total_bits": 360997440,
2646
  "q_proj": {
2647
  "group_size": {
2648
  "2": 64
@@ -2717,10 +2717,10 @@
2717
  },
2718
  "down_proj": {
2719
  "group_size": {
2720
- "2": 64
2721
  },
2722
  "bits": [
2723
- 2
2724
  ],
2725
  "bits_prop": [
2726
  1
@@ -2729,8 +2729,8 @@
2729
  }
2730
  },
2731
  "model.layers.31": {
2732
- "accuracy": 0.9416450262069702,
2733
- "total_bits": 660324384,
2734
  "q_proj": {
2735
  "group_size": {
2736
  "2": 64
@@ -2745,10 +2745,10 @@
2745
  },
2746
  "k_proj": {
2747
  "group_size": {
2748
- "2": 64
2749
  },
2750
  "bits": [
2751
- 2
2752
  ],
2753
  "bits_prop": [
2754
  1
 
89
  }
90
  },
91
  "model.layers.1": {
92
+ "accuracy": 0.7773141860961914,
93
+ "total_bits": 365124672,
94
  "q_proj": {
95
  "group_size": {
96
  "2": 64
 
129
  },
130
  "o_proj": {
131
  "group_size": {
132
+ "2": 64
133
  },
134
  "bits": [
135
+ 2
136
  ],
137
  "bits_prop": [
138
  1
 
165
  },
166
  "down_proj": {
167
  "group_size": {
168
+ "2": 64
169
  },
170
  "bits": [
171
+ 2
172
  ],
173
  "bits_prop": [
174
  1
 
265
  }
266
  },
267
  "model.layers.3": {
268
+ "accuracy": 0.8372340202331543,
269
+ "total_bits": 360997440,
270
  "q_proj": {
271
  "group_size": {
272
  "2": 64
 
341
  },
342
  "down_proj": {
343
  "group_size": {
344
+ "2": 64
345
  },
346
  "bits": [
347
+ 2
348
  ],
349
  "bits_prop": [
350
  1
 
705
  }
706
  },
707
  "model.layers.8": {
708
+ "accuracy": 0.7149877548217773,
709
+ "total_bits": 360997440,
710
  "q_proj": {
711
  "group_size": {
712
  "2": 64
 
781
  },
782
  "down_proj": {
783
  "group_size": {
784
+ "2": 64
785
  },
786
  "bits": [
787
+ 2
788
  ],
789
  "bits_prop": [
790
  1
 
793
  }
794
  },
795
  "model.layers.9": {
796
+ "accuracy": 0.9577234387397766,
797
+ "total_bits": 365124672,
798
  "q_proj": {
799
  "group_size": {
800
  "2": 64
 
809
  },
810
  "k_proj": {
811
  "group_size": {
812
+ "4": 128
813
  },
814
  "bits": [
815
+ 4
816
  ],
817
  "bits_prop": [
818
  1
 
1057
  }
1058
  },
1059
  "model.layers.12": {
1060
+ "accuracy": 0.9404451847076416,
1061
+ "total_bits": 365124672,
1062
  "q_proj": {
1063
  "group_size": {
1064
  "2": 64
 
1073
  },
1074
  "k_proj": {
1075
  "group_size": {
1076
+ "4": 128
1077
  },
1078
  "bits": [
1079
+ 4
1080
  ],
1081
  "bits_prop": [
1082
  1
 
1233
  }
1234
  },
1235
  "model.layers.14": {
1236
+ "accuracy": 0.9359749555587769,
1237
+ "total_bits": 365124672,
1238
  "q_proj": {
1239
  "group_size": {
1240
  "2": 64
 
1249
  },
1250
  "k_proj": {
1251
  "group_size": {
1252
+ "4": 128
1253
  },
1254
  "bits": [
1255
+ 4
1256
  ],
1257
  "bits_prop": [
1258
  1
 
1761
  }
1762
  },
1763
  "model.layers.20": {
1764
+ "accuracy": 0.9059760570526123,
1765
+ "total_bits": 365124672,
1766
  "q_proj": {
1767
  "group_size": {
1768
  "2": 64
 
1777
  },
1778
  "k_proj": {
1779
  "group_size": {
1780
+ "4": 128
1781
  },
1782
  "bits": [
1783
+ 4
1784
  ],
1785
  "bits_prop": [
1786
  1
 
1849
  }
1850
  },
1851
  "model.layers.21": {
1852
+ "accuracy": 0.9023456573486328,
1853
+ "total_bits": 365124672,
1854
  "q_proj": {
1855
  "group_size": {
1856
  "2": 64
 
1865
  },
1866
  "k_proj": {
1867
  "group_size": {
1868
+ "4": 128
1869
  },
1870
  "bits": [
1871
+ 4
1872
  ],
1873
  "bits_prop": [
1874
  1
 
2289
  }
2290
  },
2291
  "model.layers.26": {
2292
+ "accuracy": 0.9034652709960938,
2293
+ "total_bits": 365124672,
2294
  "q_proj": {
2295
  "group_size": {
2296
  "2": 64
 
2305
  },
2306
  "k_proj": {
2307
  "group_size": {
2308
+ "4": 128
2309
  },
2310
  "bits": [
2311
+ 4
2312
  ],
2313
  "bits_prop": [
2314
  1
 
2465
  }
2466
  },
2467
  "model.layers.28": {
2468
+ "accuracy": 0.9271254539489746,
2469
+ "total_bits": 449761824,
2470
  "q_proj": {
2471
  "group_size": {
2472
  "2": 64
 
2541
  },
2542
  "down_proj": {
2543
  "group_size": {
2544
+ "4": 128
2545
  },
2546
  "bits": [
2547
+ 4
2548
  ],
2549
  "bits_prop": [
2550
  1
 
2553
  }
2554
  },
2555
  "model.layers.29": {
2556
+ "accuracy": 0.9337887763977051,
2557
+ "total_bits": 449761824,
2558
  "q_proj": {
2559
  "group_size": {
2560
  "2": 64
 
2629
  },
2630
  "down_proj": {
2631
  "group_size": {
2632
+ "4": 128
2633
  },
2634
  "bits": [
2635
+ 4
2636
  ],
2637
  "bits_prop": [
2638
  1
 
2641
  }
2642
  },
2643
  "model.layers.30": {
2644
+ "accuracy": 0.9370880126953125,
2645
+ "total_bits": 449761824,
2646
  "q_proj": {
2647
  "group_size": {
2648
  "2": 64
 
2717
  },
2718
  "down_proj": {
2719
  "group_size": {
2720
+ "4": 128
2721
  },
2722
  "bits": [
2723
+ 4
2724
  ],
2725
  "bits_prop": [
2726
  1
 
2729
  }
2730
  },
2731
  "model.layers.31": {
2732
+ "accuracy": 0.9427725076675415,
2733
+ "total_bits": 664451616,
2734
  "q_proj": {
2735
  "group_size": {
2736
  "2": 64
 
2745
  },
2746
  "k_proj": {
2747
  "group_size": {
2748
+ "4": 128
2749
  },
2750
  "bits": [
2751
+ 4
2752
  ],
2753
  "bits_prop": [
2754
  1