moelanoby commited on
Commit
d7ca303
·
verified ·
1 Parent(s): 176871f

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +40 -0
README.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ To load this model from the Hugging Face Hub in Python:
2
+ ```python
3
+ # 0. If in Colab and it's a new session, or if model is private, authenticate:
4
+ # from huggingface_hub import notebook_login; notebook_login()
5
+
6
+ # 1. Import necessary libraries:
7
+ from transformers import AutoModelForCausalLM, AutoTokenizer
8
+ # The following torch imports might be needed if you were to define the classes manually,
9
+ # but trust_remote_code=True should handle it by loading them from the Hub.
10
+ # import torch
11
+ # import torch.nn as nn
12
+
13
+ # 2. Define your model ID:
14
+ MODEL_ID = "moelanoby/Sensitive-Qwen-0.5B"
15
+
16
+ # 3. Load tokenizer and model (trust_remote_code=True is CRUCIAL):
17
+ # This allows Transformers to download and use the Python file ('LLMadd.py')
18
+ # from your Hub repository, which contains the definitions for
19
+ # `SensitivityModule` and `SensitiveBottleneckLayer`.
20
+
21
+ try:
22
+ tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
23
+ model = AutoModelForCausalLM.from_pretrained(MODEL_ID, trust_remote_code=True, device_map='auto') # Add other params as needed
24
+ print(f'Model {MODEL_ID} loaded successfully!')
25
+ except Exception as e:
26
+ print(f'Error loading model: {e}')
27
+ print('Ensure the custom code file (LLMadd.py) in the Hub repo is correct and classes are defined.')
28
+
29
+ # 4. Example generation (adjust based on your model's chat template, e.g., Qwen2-Instruct):
30
+ # prompt = "What is the capital of France?"
31
+ # messages = [{"role": "user", "content": prompt}]
32
+ # text_input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
33
+ # model_inputs = tokenizer([text_input], return_tensors="pt").to(model.device)
34
+ # generated_ids = model.generate(model_inputs.input_ids, max_new_tokens=50)
35
+ # result = tokenizer.batch_decode(generated_ids[:, model_inputs.input_ids.shape[-1]:], skip_special_tokens=True)[0]
36
+ # print(f'Generated: {result}')
37
+ ```
38
+
39
+ IMPORTANT: `trust_remote_code=True` allows the execution of Python code
40
+ from the 'moelanoby/Sensitive-Qwen-0.5B' repository on Hugging Face Hub.