Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: malhajar/phi-2-chat
|
3 |
+
datasets:
|
4 |
+
- yahma/alpaca-cleaned
|
5 |
+
inference: false
|
6 |
+
language:
|
7 |
+
- en
|
8 |
+
model_creator: malhajar
|
9 |
+
model_name: phi-2-chat
|
10 |
+
pipeline_tag: text-generation
|
11 |
+
quantized_by: afrideva
|
12 |
+
tags:
|
13 |
+
- gguf
|
14 |
+
- ggml
|
15 |
+
- quantized
|
16 |
+
- q2_k
|
17 |
+
- q3_k_m
|
18 |
+
- q4_k_m
|
19 |
+
- q5_k_m
|
20 |
+
- q6_k
|
21 |
+
- q8_0
|
22 |
+
---
|
23 |
+
# malhajar/phi-2-chat-GGUF
|
24 |
+
|
25 |
+
Quantized GGUF model files for [phi-2-chat](https://huggingface.co/malhajar/phi-2-chat) from [malhajar](https://huggingface.co/malhajar)
|
26 |
+
|
27 |
+
|
28 |
+
| Name | Quant method | Size |
|
29 |
+
| ---- | ---- | ---- |
|
30 |
+
| [phi-2-chat.fp16.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.fp16.gguf) | fp16 | 5.56 GB |
|
31 |
+
| [phi-2-chat.q2_k.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q2_k.gguf) | q2_k | 1.17 GB |
|
32 |
+
| [phi-2-chat.q3_k_m.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q3_k_m.gguf) | q3_k_m | 1.48 GB |
|
33 |
+
| [phi-2-chat.q4_k_m.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q4_k_m.gguf) | q4_k_m | 1.79 GB |
|
34 |
+
| [phi-2-chat.q5_k_m.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q5_k_m.gguf) | q5_k_m | 2.07 GB |
|
35 |
+
| [phi-2-chat.q6_k.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q6_k.gguf) | q6_k | 2.29 GB |
|
36 |
+
| [phi-2-chat.q8_0.gguf](https://huggingface.co/afrideva/phi-2-chat-GGUF/resolve/main/phi-2-chat.q8_0.gguf) | q8_0 | 2.96 GB |
|
37 |
+
|
38 |
+
|
39 |
+
|
40 |
+
## Original Model Card:
|
41 |
+
# Model Card for Model ID
|
42 |
+
|
43 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
44 |
+
malhajar/phi-2-chat is a finetuned version of [`phi-2`]( https://huggingface.co/microsoft/phi-2) using SFT Training.
|
45 |
+
This model can answer information in a chat format as it is finetuned specifically on instructions specifically [`alpaca-cleaned`]( https://huggingface.co/datasets/yahma/alpaca-cleaned)
|
46 |
+
|
47 |
+
### Model Description
|
48 |
+
|
49 |
+
- **Developed by:** [`Mohamad Alhajar`](https://www.linkedin.com/in/muhammet-alhajar/)
|
50 |
+
- **Language(s) (NLP):** Turkish
|
51 |
+
- **Finetuned from model:** [`microsoft/phi-2`](https://huggingface.co/microsoft/phi-2)
|
52 |
+
|
53 |
+
### Prompt Template
|
54 |
+
```
|
55 |
+
### Instruction:
|
56 |
+
|
57 |
+
<prompt> (without the <>)
|
58 |
+
|
59 |
+
### Response:
|
60 |
+
```
|
61 |
+
|
62 |
+
## How to Get Started with the Model
|
63 |
+
|
64 |
+
Use the code sample provided in the original post to interact with the model.
|
65 |
+
```python
|
66 |
+
from transformers import AutoTokenizer,AutoModelForCausalLM
|
67 |
+
|
68 |
+
model_id = "malhajar/phi-2-chat"
|
69 |
+
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
|
70 |
+
device_map="auto",
|
71 |
+
torch_dtype=torch.float16,
|
72 |
+
revision="main")
|
73 |
+
|
74 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
75 |
+
|
76 |
+
question: "Türkiyenin en büyük şehir nedir?"
|
77 |
+
# For generating a response
|
78 |
+
prompt = '''
|
79 |
+
### Instruction: {question} ### Response:
|
80 |
+
'''
|
81 |
+
input_ids = tokenizer(prompt, return_tensors="pt").input_ids
|
82 |
+
output = model.generate(inputs=input_ids,max_new_tokens=512,pad_token_id=tokenizer.eos_token_id,top_k=50, do_sample=True,repetition_penalty=1.3
|
83 |
+
top_p=0.95,trust_remote_code=True,)
|
84 |
+
response = tokenizer.decode(output[0])
|
85 |
+
|
86 |
+
print(response)
|
87 |
+
```
|