Update README.md
Browse files
README.md
CHANGED
@@ -10,10 +10,14 @@ base_model:
|
|
10 |
- mlabonne/ChimeraLlama-3-8B-v2
|
11 |
- nbeerbower/llama-3-stella-8B
|
12 |
- uygarkurt/llama-3-merged-linear
|
|
|
13 |
---
|
14 |
|
15 |
# NeuralLLaMa-3-8b-DT-v0.1
|
16 |
|
|
|
|
|
|
|
17 |
NeuralLLaMa-3-8b-DT-v0.1 is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
18 |
* [mlabonne/ChimeraLlama-3-8B-v2](https://huggingface.co/mlabonne/ChimeraLlama-3-8B-v2)
|
19 |
* [nbeerbower/llama-3-stella-8B](https://huggingface.co/nbeerbower/llama-3-stella-8B)
|
@@ -43,28 +47,43 @@ parameters:
|
|
43 |
int8_mask: true
|
44 |
dtype: float16
|
45 |
```
|
|
|
|
|
|
|
46 |
|
47 |
## 💻 Usage
|
48 |
|
49 |
```python
|
50 |
-
!pip install -qU transformers accelerate
|
51 |
|
52 |
-
from transformers import AutoTokenizer
|
53 |
-
import transformers
|
54 |
import torch
|
55 |
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
pipeline = transformers.pipeline(
|
62 |
-
"text-generation",
|
63 |
-
model=model,
|
64 |
-
torch_dtype=torch.float16,
|
65 |
-
device_map="auto",
|
66 |
)
|
67 |
|
68 |
-
|
69 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
```
|
|
|
10 |
- mlabonne/ChimeraLlama-3-8B-v2
|
11 |
- nbeerbower/llama-3-stella-8B
|
12 |
- uygarkurt/llama-3-merged-linear
|
13 |
+
license: other
|
14 |
---
|
15 |
|
16 |
# NeuralLLaMa-3-8b-DT-v0.1
|
17 |
|
18 |
+
|
19 |
+

|
20 |
+
|
21 |
NeuralLLaMa-3-8b-DT-v0.1 is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
22 |
* [mlabonne/ChimeraLlama-3-8B-v2](https://huggingface.co/mlabonne/ChimeraLlama-3-8B-v2)
|
23 |
* [nbeerbower/llama-3-stella-8B](https://huggingface.co/nbeerbower/llama-3-stella-8B)
|
|
|
47 |
int8_mask: true
|
48 |
dtype: float16
|
49 |
```
|
50 |
+
## 🗨️ Chats
|
51 |
+
|
52 |
+

|
53 |
|
54 |
## 💻 Usage
|
55 |
|
56 |
```python
|
57 |
+
!pip install -qU transformers accelerate bitsandbytes
|
58 |
|
59 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer, BitsAndBytesConfig
|
|
|
60 |
import torch
|
61 |
|
62 |
+
bnb_config = BitsAndBytesConfig(
|
63 |
+
load_in_4bit=True,
|
64 |
+
bnb_4bit_use_double_quant=True,
|
65 |
+
bnb_4bit_quant_type="nf4",
|
66 |
+
bnb_4bit_compute_dtype=torch.bfloat16
|
|
|
|
|
|
|
|
|
|
|
67 |
)
|
68 |
|
69 |
+
MODEL_NAME = 'Kukedlc/NeuralLLaMa-3-8b-DT-v0.1'
|
70 |
+
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
|
71 |
+
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, device_map='cuda:0', quantization_config=bnb_config)
|
72 |
+
|
73 |
+
prompt_system = "You are an advanced language model that speaks Spanish fluently, clearly, and precisely.\
|
74 |
+
You are called Roberto the Robot and you are an aspiring post-modern artist."
|
75 |
+
prompt = "Create a piece of art that represents how you see yourself, Roberto, as an advanced LLm, with ASCII art, mixing diagrams, engineering and let yourself go."
|
76 |
+
|
77 |
+
chat = [
|
78 |
+
{"role": "system", "content": f"{prompt_system}"},
|
79 |
+
{"role": "user", "content": f"{prompt}"},
|
80 |
+
]
|
81 |
+
|
82 |
+
chat = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
|
83 |
+
inputs = tokenizer(chat, return_tensors="pt").to('cuda')
|
84 |
+
streamer = TextStreamer(tokenizer)
|
85 |
+
stop_token = "<|eot_id|>"
|
86 |
+
stop = tokenizer.encode(stop_token)[0]
|
87 |
+
|
88 |
+
_ = model.generate(**inputs, streamer=streamer, max_new_tokens=1024, do_sample=True, temperature=0.7, repetition_penalty=1.2, top_p=0.9, eos_token_id=stop)
|
89 |
```
|