Featherless
Mistral-7B-Instruct-v0.2-attention-sparsity-20Wang7776
7B Params FP8 Open Weights

wang7776/Mistral-7B-Instruct-v0.2-attention-sparsity-20 is a 7 billion parameter instruction-tuned language model, based on Mistral-7B-Instruct-v0.2, that has been pruned to 20% attention sparsity using the Wanda method. This pruning technique reduces model size and computational requirements without retraining, making it efficient for deployment. It retains the original model's capabilities for instruction-following tasks, leveraging Grouped-Query Attention and Sliding-Window Attention.

Loading preview...

Quick Stats

115

Parameters:7BContext length:8kArchitecture:TransformerPrecision:FP8Quantized variants:AvailableLast updated:January 2024
0.0M

Model tree for

wang7776/Mistral-7B-Instruct-v0.2-attention-sparsity-20
Popular Sampler Settings

Most commonly used values from Featherless users

temperature

This setting influences the sampling randomness. Lower values make the model more deterministic; higher values introduce randomness. Zero is greedy sampling.

0.4

top_p

This setting controls the cumulative probability of considered top tokens. Must be in (0, 1]. Set to 1 to consider all tokens.

0.3

top_k

This limits the number of top tokens to consider. Set to -1 to consider all tokens.

50

frequency_penalty

This setting penalizes new tokens based on their frequency in the generated text. Values > 0 encourage new tokens; < 0 encourages repetition.

0.5

presence_penalty

This setting penalizes new tokens based on their presence in the generated text so far. Values > 0 encourage new tokens; < 0 encourages repetition.

0.5

repetition_penalty

This setting penalizes new tokens based on their appearance in the prompt and generated text. Values > 1 encourage new tokens; < 1 encourages repetition.

1.1

min_p

This setting representing the minimum probability for a token to be considered relative to the most likely token. Must be in [0, 1]. Set to 0 to disable.

0.05