Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,45 @@
|
|
| 1 |
-
---
|
| 2 |
license: apache-2.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
license: apache-2.0
|
| 2 |
---
|
| 3 |
+
# GreenBit Yi
|
| 4 |
+
|
| 5 |
+
This is GreenBitAI's pretrained **2-bit** Yi 34B model with extreme compression yet still strong performance.
|
| 6 |
+
|
| 7 |
+
Please refer to our [Github page](https://github.com/GreenBitAI/low_bit_llama) for the code to run the model and more information.
|
| 8 |
+
|
| 9 |
+
## Model Description
|
| 10 |
+
|
| 11 |
+
- **Developed by:** [GreenBitAI](https://github.com/GreenBitAI)
|
| 12 |
+
- **Model type:** Causal (Llama 2/Yi 34B)
|
| 13 |
+
- **Language(s) (NLP):** English
|
| 14 |
+
- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0), [Llama 2 license agreement](https://ai.meta.com/resources/models-and-libraries/llama-downloads/)
|
| 15 |
+
|
| 16 |
+
| Model | Yi-34B | | | Yi-6B | | |
|
| 17 |
+
| --- | --- | --- | --- | --- | --- | --- |
|
| 18 |
+
| Bit | 16 | 4 | 2 | 16 | 4 | 2 |
|
| 19 |
+
| GroupSize | - | 32 | 8 | - | 32 | 8 |
|
| 20 |
+
| Model Size (GB) | 68.79 | 19.89 | 15.508 | 12.12 | 4.04 | 3.32 |
|
| 21 |
+
| AVG | 70.64 | 69.7 | 65.78 | 60.11 | 59.14 | 53.69 |
|
| 22 |
+
| Detailed Evaluation | | | | | | |
|
| 23 |
+
| MMLU | 76.32 | 75.42 | 72 | 63.24 | 62.09 | 58.57 |
|
| 24 |
+
| CMMLU | 83.65 | 83.07 | 78.26 | 75.53 | 72.85 | 64.83 |
|
| 25 |
+
| ARC-e | 84.42 | 84.13 | 81.35 | 77.23 | 76.52 | 73.15 |
|
| 26 |
+
| ARC-c | 61.77 | 59.56 | 57 | 50.34 | 48.47 | 42.58 |
|
| 27 |
+
| GAOKAO | 82.8 | 81.37 | 77.9 | 72.2 | 72.87 | 64.67 |
|
| 28 |
+
| GSM8K | 67.24 | 63.61 | 50.1895 | 32.52 | 28.05 | 16.98 |
|
| 29 |
+
| HumanEval | 25.6 | 25 | 18.9024 | 15.85 | 15.85 | 12.19 |
|
| 30 |
+
| BBH | 54.3 | 52.3 | 49.0401 | 42.8 | 41.47 | 38.12 |
|
| 31 |
+
| WinoGrande | 78.68 | 78.53 | 77.98 | 70.63 | 71.19 | 67.96 |
|
| 32 |
+
| PIQA | 82.86 | 82.75 | 81.72 | 78.56 | 79.05 | 76.55 |
|
| 33 |
+
| SIQA | 74.46 | 73.44 | 71.6 | 64.53 | 64.53 | 57 |
|
| 34 |
+
| HellaSwag | 83.64 | 83.02 | 80.7 | 74.91 | 73.27 | 69.07 |
|
| 35 |
+
| OBQA | 91.6 | 90.8 | 84 | 85.4 | 82.6 | 79.4 |
|
| 36 |
+
| CSQA | 83.37 | 83.05 | 81.5725 | 76.9 | 75.43 | 69.53 |
|
| 37 |
+
| TriviaQA | 81.52 | 80.73 | 75.8163 | 64.85 | 61.75 | 50.5 |
|
| 38 |
+
| SquAD | 92.46 | 91.12 | 89.3624 | 88.95 | 88.39 | 84.22 |
|
| 39 |
+
| BoolQ | 88.25 | 88.17 | 80.74 | 76.23 | 77.1 | 75.5 |
|
| 40 |
+
| MBPP | 41 | 39.68 | 31.4815 | 26.32 | 25.13 | 16 |
|
| 41 |
+
| QUAC | 48.61 | 47.43 | 46.5406 | 40.92 | 40.16 | 37.83 |
|
| 42 |
+
| Lambda | 73.18 | 73.39 | 72.83 | 67.74 | 67.8 | 61.8 |
|
| 43 |
+
| NaturalQuestion | 27.67 | 27.21 | 22.4994 | 16.69 | 17.42 | 11.05|
|
| 44 |
+
|
| 45 |
+
|