RichardErkhov/McGill-NLP_-_Sheared-LLaMA-2.7B-weblinx-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Sheared-LLaMA-2.7B-weblinx - GGUF

• Model creator: https://huggingface.co/McGill-NLP/
• Original model: https://huggingface.co/McGill-NLP/Sheared-LLaMA-2.7B-weblinx/

Name	Quant method	Size
Sheared-LLaMA-2.7B-weblinx.Q2_K.gguf	Q2_K	0.96GB
Sheared-LLaMA-2.7B-weblinx.IQ3_XS.gguf	IQ3_XS	1.06GB
Sheared-LLaMA-2.7B-weblinx.IQ3_S.gguf	IQ3_S	1.11GB
Sheared-LLaMA-2.7B-weblinx.Q3_K_S.gguf	Q3_K_S	1.11GB
Sheared-LLaMA-2.7B-weblinx.IQ3_M.gguf	IQ3_M	1.17GB
Sheared-LLaMA-2.7B-weblinx.Q3_K.gguf	Q3_K	1.24GB
Sheared-LLaMA-2.7B-weblinx.Q3_K_M.gguf	Q3_K_M	1.24GB
Sheared-LLaMA-2.7B-weblinx.Q3_K_L.gguf	Q3_K_L	1.35GB
Sheared-LLaMA-2.7B-weblinx.IQ4_XS.gguf	IQ4_XS	1.37GB
Sheared-LLaMA-2.7B-weblinx.Q4_0.gguf	Q4_0	1.44GB
Sheared-LLaMA-2.7B-weblinx.IQ4_NL.gguf	IQ4_NL	1.44GB
Sheared-LLaMA-2.7B-weblinx.Q4_K_S.gguf	Q4_K_S	1.45GB
Sheared-LLaMA-2.7B-weblinx.Q4_K.gguf	Q4_K	1.53GB
Sheared-LLaMA-2.7B-weblinx.Q4_K_M.gguf	Q4_K_M	1.53GB
Sheared-LLaMA-2.7B-weblinx.Q4_1.gguf	Q4_1	1.59GB
Sheared-LLaMA-2.7B-weblinx.Q5_0.gguf	Q5_0	1.74GB
Sheared-LLaMA-2.7B-weblinx.Q5_K_S.gguf	Q5_K_S	1.74GB
Sheared-LLaMA-2.7B-weblinx.Q5_K.gguf	Q5_K	1.79GB
Sheared-LLaMA-2.7B-weblinx.Q5_K_M.gguf	Q5_K_M	1.79GB
Sheared-LLaMA-2.7B-weblinx.Q5_1.gguf	Q5_1	1.89GB
Sheared-LLaMA-2.7B-weblinx.Q6_K.gguf	Q6_K	2.07GB
Sheared-LLaMA-2.7B-weblinx.Q8_0.gguf	Q8_0	2.67GB

Original model description:

datasets: - McGill-NLP/WebLINX - McGill-NLP/WebLINX-full language: - en metrics: - f1 - iou - chrf library_name: transformers pipeline_tag: text-generation tags: - weblinx - text-generation-inference - web-agents - agents license: llama2

WebLINX: Real-World Website Navigation with Multi-Turn Dialogue

Xing Han Lù*, Zdeněk Kasner*, Siva Reddy

📄Paper

🌐Website

📓Colab

🤗Dataset

💾Code

Quickstart

from datasets import load_dataset
from huggingface_hub import snapshot_download
from transformers import pipeline

# Load validation split
valid = load_dataset("McGill-NLP/weblinx", split="validation")

# Download and load the templates
snapshot_download(
    "McGill-NLP/WebLINX", repo_type="dataset", allow_patterns="templates/*.txt", local_dir="./"
)
with open('templates/llama.txt') as f:
    template = f.read()

turn = valid[0]
turn_text = template.format(**turn)

# Load action model and input the text to get prediction
action_model = pipeline(
    model="McGill-NLP/Sheared-LLaMA-2.7B-weblinx", device=0, torch_dtype='auto'
)
out = action_model(turn_text, return_full_text=False, max_new_tokens=64, truncation=True)
pred = out[0]['generated_text']

print("Ref:", turn["action"])
print("Pred:", pred)

Original Model

This model is finetuned on WebLINX using checkpoints previously published on Huggingface Hub.
Click here to access the original model.

License

This model is derived from LLaMA-2, which can only be used with the LLaMA 2 Community License Agreement. By using or distributing any portion or element of this model, you agree to be bound by this Agreement.