Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
## Perplexity (ctx 8192)
|
| 2 |
+
|
| 3 |
+
BF16:
|
| 4 |
+
```
|
| 5 |
+
I1005 10:12:54.337877 967332 eval_ppl.py:66] wikitext2 perplexity: 6.49992036819458
|
| 6 |
+
I1005 10:17:19.718732 967332 eval_ppl.py:66] c4 perplexity: 8.022844314575195
|
| 7 |
+
```
|
| 8 |
+
2 Bit QTIP (this model):
|
| 9 |
+
```
|
| 10 |
+
I1005 10:13:20.523440 967333 eval_ppl.py:66] wikitext2 perplexity: 7.8186869621276855
|
| 11 |
+
I1005 10:19:34.002755 967333 eval_ppl.py:66] c4 perplexity: 9.199813842773438
|
| 12 |
+
```
|
| 13 |
+
|
| 14 |
+
## Zeroshot Results with lm_eval 0.3.0
|
| 15 |
+
BF16:
|
| 16 |
+
```
|
| 17 |
+
|-------------|------:|--------|-----:|---|-----:|
|
| 18 |
+
|arc_challenge| 0|acc |0.5196|± |0.0146|
|
| 19 |
+
| | |acc_norm|0.5512|± |0.0145|
|
| 20 |
+
|arc_easy | 0|acc |0.8182|± |0.0079|
|
| 21 |
+
| | |acc_norm|0.7980|± |0.0082|
|
| 22 |
+
|boolq | 1|acc |0.8410|± |0.0064|
|
| 23 |
+
|piqa | 0|acc |0.8003|± |0.0093|
|
| 24 |
+
| | |acc_norm|0.8090|± |0.0092|
|
| 25 |
+
|winogrande | 0|acc |0.7380|± |0.0124|
|
| 26 |
+
```
|
| 27 |
+
2 Bit QTIP (this model):
|
| 28 |
+
```
|
| 29 |
+
|-------------|------:|--------|-----:|---|-----:|
|
| 30 |
+
|arc_challenge| 0|acc |0.4676|± |0.0146|
|
| 31 |
+
| | |acc_norm|0.5077|± |0.0146|
|
| 32 |
+
|arc_easy | 0|acc |0.7769|± |0.0085|
|
| 33 |
+
| | |acc_norm|0.7551|± |0.0088|
|
| 34 |
+
|boolq | 1|acc |0.8101|± |0.0069|
|
| 35 |
+
|piqa | 0|acc |0.7775|± |0.0097|
|
| 36 |
+
| | |acc_norm|0.7829|± |0.0096|
|
| 37 |
+
|winogrande | 0|acc |0.6985|± |0.0129|
|
| 38 |
+
```
|