Quantization made by Richard Erkhov.

TinyLlama-1.1B-32k-Instruct - GGUF

Model creator: https://huggingface.co/Doctor-Shotgun/
Original model: https://huggingface.co/Doctor-Shotgun/TinyLlama-1.1B-32k-Instruct/

Name	Quant method	Size
TinyLlama-1.1B-32k-Instruct.Q2_K.gguf	Q2_K	0.4GB
TinyLlama-1.1B-32k-Instruct.IQ3_XS.gguf	IQ3_XS	0.44GB
TinyLlama-1.1B-32k-Instruct.IQ3_S.gguf	IQ3_S	0.47GB
TinyLlama-1.1B-32k-Instruct.Q3_K_S.gguf	Q3_K_S	0.47GB
TinyLlama-1.1B-32k-Instruct.IQ3_M.gguf	IQ3_M	0.48GB
TinyLlama-1.1B-32k-Instruct.Q3_K.gguf	Q3_K	0.51GB
TinyLlama-1.1B-32k-Instruct.Q3_K_M.gguf	Q3_K_M	0.51GB
TinyLlama-1.1B-32k-Instruct.Q3_K_L.gguf	Q3_K_L	0.55GB
TinyLlama-1.1B-32k-Instruct.IQ4_XS.gguf	IQ4_XS	0.57GB
TinyLlama-1.1B-32k-Instruct.Q4_0.gguf	Q4_0	0.59GB
TinyLlama-1.1B-32k-Instruct.IQ4_NL.gguf	IQ4_NL	0.6GB
TinyLlama-1.1B-32k-Instruct.Q4_K_S.gguf	Q4_K_S	0.6GB
TinyLlama-1.1B-32k-Instruct.Q4_K.gguf	Q4_K	0.62GB
TinyLlama-1.1B-32k-Instruct.Q4_K_M.gguf	Q4_K_M	0.62GB
TinyLlama-1.1B-32k-Instruct.Q4_1.gguf	Q4_1	0.65GB
TinyLlama-1.1B-32k-Instruct.Q5_0.gguf	Q5_0	0.71GB
TinyLlama-1.1B-32k-Instruct.Q5_K_S.gguf	Q5_K_S	0.71GB
TinyLlama-1.1B-32k-Instruct.Q5_K.gguf	Q5_K	0.73GB
TinyLlama-1.1B-32k-Instruct.Q5_K_M.gguf	Q5_K_M	0.73GB
TinyLlama-1.1B-32k-Instruct.Q5_1.gguf	Q5_1	0.77GB
TinyLlama-1.1B-32k-Instruct.Q6_K.gguf	Q6_K	0.84GB
TinyLlama-1.1B-32k-Instruct.Q8_0.gguf	Q8_0	1.09GB

Original model description:

inference: false language: - en library_name: transformers pipeline_tag: text-generation tags: - llama datasets: - LDJnr/Capybara - jondurbin/airoboros-3.2 - unalignment/toxic-dpo-v0.1 - LDJnr/Verified-Camel - HuggingFaceH4/no_robots - Doctor-Shotgun/no-robots-sharegpt - Doctor-Shotgun/capybara-sharegpt

Norobara-ZLoss-8x7B

This is an instruct-tuned TinyLlama-1.1B-32k on several open-source instruct datasets, intended primarily for speculative decoding.

Usage:

The intended prompt format is a modified multi-turn Alpaca instruction format:

### Instruction:
{system prompt}

### Input:
{user message}

### Response:
{model response}

### Input:
{user message}

### Response:
{model response}

(etc.)

Bias, Risks, and Limitations

The model will show biases present in the base model. No ethical alignment was applied to prevent the generation of toxic or harmful outputs (in fact the opposite, with examples from toxic-DPO included), so generate at your own risk.

Training Details

This model was trained as a full finetune for 3 epochs using a single A100 GPU for around 3.5 hours.