Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
GGUF format files of the model vinai/PhoGPT-4B-Chat.
|
| 2 |
+
|
| 3 |
+
I'm trying to get PhoGPT to work with llama-cpp and llama-cpp-python.
|
| 4 |
+
|
| 5 |
+
I cannot get [nguyenviet/PhoGPT-4B-Chat-GGUF](https://huggingface.co/nguyenviet/PhoGPT-4B-Chat-GGUF) to work in Colab:
|
| 6 |
+
|
| 7 |
+
```
|
| 8 |
+
from llama_cpp import Llama
|
| 9 |
+
|
| 10 |
+
llm = Llama.from_pretrained(
|
| 11 |
+
repo_id="nguyenviet/PhoGPT-4B-Chat-GGUF",
|
| 12 |
+
filename="*q3_k_m.gguf*",
|
| 13 |
+
)
|
| 14 |
+
|
| 15 |
+
...
|
| 16 |
+
llama_model_load: error loading model: done_getting_tensors: wrong number of tensors; expected 388, got 387
|
| 17 |
+
llama_load_model_from_file: failed to load model
|
| 18 |
+
...
|
| 19 |
+
```
|
| 20 |
+
|
| 21 |
+
My [issue](https://github.com/VinAIResearch/PhoGPT/issues/22) was resolved (thanks to @nviet and @datquocnguyen), and I figure people want to try the model in Colab. So I created my own `GGUF` file.
|