README.md

metadata

library_name: transformers
tags: []
widget:
  - example_title: EMO 1
    messages:
      - role: system
        content: >-
          You are a helpful and emotional assistant that will always respond in
          EMO style.
      - role: user
        content: >-
          Imagine you're helping someone who is feeling overhelmed. How do you
          feel in this situation?
  - example_title: EMO 2
    messages:
      - role: system
        content: >-
          You are a helpful and emotional assistant that will always respond in
          EMO style.
      - role: user
        content: >-
          My best friend recently lost their parent to cancer after a long
          battle. They are understandably devastated and struggling with grief.
inference:
  parameters:
    max_new_tokens: 1024
    do_sample: true

Model card comming soon

from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    "Abhaykoul/EMO-1B",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Abhaykoul/EMO-1B")

prompt = "Imagine you're helping someone who is feeling overwhelmed. How do you feel in this situation?"
messages = [
    {"role": "system", "content": "You are a helpful and emotional assistant that will always respond in EMO style"},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(device)

generated_ids = model.generate(
    model_inputs.input_ids,
    max_new_tokens=512
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)