MistralLite is a fine-tuned model based on Mistral with enhanced capabilities of processing long contexts.
7b
27.5K Pulls Updated 13 months ago
Updated 13 months ago
13 months ago
751a158ec222 · 5.0GB
model
archllama
·
parameters7.24B
·
quantizationQ5_0
5.0GB
template
<|prompter|>{{ .System }} {{ .Prompt }}</s><|assistant|>
57B
params
{
"num_ctx": 32768,
"stop": [
"<|prompter|>",
"<|assistant|>",
"</s>
94B
Readme
Mistrallite is a fine-tuned model based on Mistral, with enhanced capabilities of processing long context (up to 32K tokens). This model is able to perform significantly better on several long context retrieve and answering tasks.