Update README.md

README.md

@@ -4,7 +4,7 @@ base_model: EleutherAI/pythia-2.8b-deduped
 tags:
 - generated_from_trainer
 model-index:
-- name: PythiaChat-2.7B
+- name: PythiaChat-2.8B_v0.1
   results: []
 library_name: peft
 ---
@@ -12,23 +12,57 @@ library_name: peft
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-# PythiaChat-2.7B
+# PythiaChat-2.8B_v0.1
 
-This model is a fine-tuned version of [EleutherAI/pythia-2.8b-deduped](https://huggingface.co/EleutherAI/pythia-2.8b-deduped) on an unknown dataset.
+This model is a fine-tuned version of [EleutherAI/pythia-2.8b-deduped](https://huggingface.co/EleutherAI/pythia-2.8b-deduped) on the [Baize dataset](https://huggingface.co/datasets/linkanjarad/baize-chat-data/viewer/linkanjarad--baize-chat-data), trained for only 200+ steps for testing. This model is trained for "chat" style instruction following capabilities.
 
-## Model description
+# Sample Use
 
-More information needed
+Remember to mark the human messages with `[|Human|]` and AI messages with `[|AI]` at the start.
 
-## Intended uses & limitations
 
-More information needed
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from peft import PeftModel, PeftConfig
 
-## Training and evaluation data
+peft_model_id = "linkanjarad/PythiaChat-2.8B_v0.1"
+model_id = "EleutherAI/pythia-2.8b-deduped"
+model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", trust_remote_code=True) # you can add `load_in_4bit=True` for faster inference
+model = PeftModel.from_pretrained(model, peft_model_id)
 
-More information needed
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = model.to('cuda')
+model.eval()
 
-## Training procedure
+
+input_text = """The conversation between human and AI assistant.
+[|Human|] How do I open a file with python?
+[|AI|]"""
+
+# Tokenize the input text
+input_ids = tokenizer.encode(input_text, return_tensors='pt').to('cuda')
+len_input = len(input_ids[0])
+# Generate text using the model
+with torch.no_grad():
+    output = model.generate(input_ids=input_ids, max_length=len_input+120, temperature=0.9, do_sample=True)
+
+# Decode the generated output
+generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
+
+print(generated_text)
+```
+
+
+Example Output
+
+```
+The conversation between human and AI assistant.
+[|Human|] How do I open a file with python?
+[|AI|] To open a file with python, you can use the open function as follows:
+
+>>> with open('filename.txt', 'w') as f:
+...     f.write(data)
+```
 
 ### Training hyperparameters