You can deploy this model and start earning money today!

ddh0/
OrcaMaid-v2-FIX-13b-32k

Name: OrcaMaid-v2-FIX-13b-32k
Rating: 2.5 (16 reviews)

text-generation-inference

endpoints_compatible

Views

No views yet

OrcaMaid-13b-v2-FIX-32k

This is the fixed version of OrcaMaid-v2-13b, extended to 32768 context length via YaRN. The (now-deleted) v2 model had issues with the merged tokenizer that prevented it from stopping when necessary, and caused it to output broken ChatML tokens like <|im_end, etc.

This is a gradient SLERP merge of Microsoft's Orca-2-13b and Undi and IkariDev's Noromaid-v0.1.1-13b, biased towards Orca.

Just as with OrcaMaid v1, the overall goal of this merge is to create a model that sounds uniquely human and natural, without sacrificing intelligence.

The prompt format is Alpaca. You can use the standard format as shown, but for best results, you should customize the system prompt to your specific needs.

Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{YOUR MESSAGE HERE}

### Response:
{BOT MESSAGE HERE}

Misc. information

BOS token is <s>
EOS token is </s>
Native context length is 32768 via YaRN (original context length was 4096)
Base model is Llama 2
Due to the inclusion of Orca-2-13b, the model is subject to the terms of the Microsoft Research License

Thanks

Thanks to Charles Goddard for his kind help with mergekit (as always)
Thanks to Undi and IkariDev for Noromaid

ddh0/OrcaMaid-v2-FIX-13b-32k

OrcaMaid-13b-v2-FIX-32k

Misc. information

Thanks

ddh0/
OrcaMaid-v2-FIX-13b-32k