Upload README.md with huggingface_hub

README.md

@@ -26,32 +26,14 @@ Chinese language model trained on pretrain dataset.
 
 ## Usage
 ```python
-import torch
-from model.model import Transformer
-from model.LMConfig import LMConfig
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
-config = LMConfig()
-model = Transformer(config)
-checkpoint = torch.load('pretrain_512.pth')
-model.load_state_dict(checkpoint['model'])
-```
-
-## Model Description
-This is a lightweight Chinese language model trained on a 4.33GB pretrain dataset. The model uses a standard transformer architecture optimized for Chinese text processing.
-
-## Intended Uses
-- Chinese text generation
-- Language modeling
-- Text completion
-- Educational purposes
+model = AutoModelForCausalLM.from_pretrained("samz/minimind-pretrain")
+tokenizer = AutoTokenizer.from_pretrained("samz/minimind-pretrain")
 
-## Limitations
-- Limited to Chinese language
-- Maximum sequence length of 1024 tokens
-- Not suitable for production workloads
-
-## Training Procedure
-- Trained on A6000 GPU
-- Learning rate: 2e-4
-- Batch size: 128
-- Training epochs: 20
+text = "今天天气真不错"
+inputs = tokenizer(text, return_tensors="pt")
+outputs = model.generate(**inputs, max_length=50)
+result = tokenizer.decode(outputs[0])
+print(result)
+```