Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: BEE-spoke-data/Mixtral-GQA-400m-v2
|
| 3 |
+
inference: false
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
license: apache-2.0
|
| 7 |
+
model_creator: BEE-spoke-data
|
| 8 |
+
model_name: Mixtral-GQA-400m-v2
|
| 9 |
+
pipeline_tag: text-generation
|
| 10 |
+
quantized_by: afrideva
|
| 11 |
+
tags:
|
| 12 |
+
- gguf
|
| 13 |
+
- ggml
|
| 14 |
+
- quantized
|
| 15 |
+
- q2_k
|
| 16 |
+
- q3_k_m
|
| 17 |
+
- q4_k_m
|
| 18 |
+
- q5_k_m
|
| 19 |
+
- q6_k
|
| 20 |
+
- q8_0
|
| 21 |
+
---
|
| 22 |
+
# BEE-spoke-data/Mixtral-GQA-400m-v2-GGUF
|
| 23 |
+
|
| 24 |
+
Quantized GGUF model files for [Mixtral-GQA-400m-v2](https://huggingface.co/BEE-spoke-data/Mixtral-GQA-400m-v2) from [BEE-spoke-data](https://huggingface.co/BEE-spoke-data)
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
| Name | Quant method | Size |
|
| 28 |
+
| ---- | ---- | ---- |
|
| 29 |
+
| [mixtral-gqa-400m-v2.fp16.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.fp16.gguf) | fp16 | 4.01 GB |
|
| 30 |
+
| [mixtral-gqa-400m-v2.q2_k.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q2_k.gguf) | q2_k | 703.28 MB |
|
| 31 |
+
| [mixtral-gqa-400m-v2.q3_k_m.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q3_k_m.gguf) | q3_k_m | 899.86 MB |
|
| 32 |
+
| [mixtral-gqa-400m-v2.q4_k_m.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q4_k_m.gguf) | q4_k_m | 1.15 GB |
|
| 33 |
+
| [mixtral-gqa-400m-v2.q5_k_m.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q5_k_m.gguf) | q5_k_m | 1.39 GB |
|
| 34 |
+
| [mixtral-gqa-400m-v2.q6_k.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q6_k.gguf) | q6_k | 1.65 GB |
|
| 35 |
+
| [mixtral-gqa-400m-v2.q8_0.gguf](https://huggingface.co/afrideva/Mixtral-GQA-400m-v2-GGUF/resolve/main/mixtral-gqa-400m-v2.q8_0.gguf) | q8_0 | 2.13 GB |
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
## Original Model Card:
|
| 40 |
+
# BEE-spoke-data/Mixtral-GQA-400m-v2
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
## testing code
|
| 46 |
+
|
| 47 |
+
```python
|
| 48 |
+
# !pip install -U -q transformers datasets accelerate sentencepiece
|
| 49 |
+
import pprint as pp
|
| 50 |
+
from transformers import pipeline
|
| 51 |
+
|
| 52 |
+
pipe = pipeline(
|
| 53 |
+
"text-generation",
|
| 54 |
+
model="BEE-spoke-data/Mixtral-GQA-400m-v2",
|
| 55 |
+
device_map="auto",
|
| 56 |
+
)
|
| 57 |
+
pipe.model.config.pad_token_id = pipe.model.config.eos_token_id
|
| 58 |
+
|
| 59 |
+
prompt = "My favorite movie is Godfather because"
|
| 60 |
+
|
| 61 |
+
res = pipe(
|
| 62 |
+
prompt,
|
| 63 |
+
max_new_tokens=256,
|
| 64 |
+
top_k=4,
|
| 65 |
+
penalty_alpha=0.6,
|
| 66 |
+
use_cache=True,
|
| 67 |
+
no_repeat_ngram_size=4,
|
| 68 |
+
repetition_penalty=1.1,
|
| 69 |
+
renormalize_logits=True,
|
| 70 |
+
)
|
| 71 |
+
pp.pprint(res[0])
|
| 72 |
+
```
|