Uploaded model
- Developed by: msfm
- License: apache-2.0
- Finetuned from model : llm-jp/llm-jp-3-13b
This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.
Example
from unsloth import FastLanguageModel
model, tokenizer = FastLanguageModel.from_pretrained(
model_name="msfm/llm-jp-3-13b-ichikara_all",
dtype=dtype,
load_in_4bit=True,
trust_remote_code=True,
)
FastLanguageModel.for_inference(model)
input = "้็้ธๆใไปใทใผใบใณๆดป่บใใใใใซๅใ็ตใในใ5ใคใฎใใจใๆใใฆใใ ใใใ"
prompt = f"""### ๆ็คบ\n{input}\n### ๅ็ญ\n"""
inputs = tokenizer([prompt], return_tensors = "pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens = 2048, use_cache = True, do_sample=False, repetition_penalty=1.2)
prediction = tokenizer.decode(outputs[0], skip_special_tokens=True).split('\n### ๅ็ญ')[-1]
Model tree for msfm/llm-jp-3-13b-ichikara_all
Base model
llm-jp/llm-jp-3-13b