neo_7b_sft_v0.1-IMat-GGUF

Llama.cpp imatrix quantization of m-a-p/neo_7b_sft_v0.1

Original Model: m-a-p/neo_7b_sft_v0.1
Original dtype: BF16 (bfloat16)
Quantized by: llama.cpp b3051
IMatrix dataset: here


Files

IMatrix

Status: βœ… Available
Link: here

Common Quants

Filename Quant type File Size Status Uses IMatrix Is Split
neo_7b_sft_v0.1.Q8_0.gguf Q8_0 8.28GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q6_K.gguf Q6_K 6.40GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q4_K.gguf Q4_K 4.74GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q3_K.gguf Q3_K 3.79GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q2_K.gguf Q2_K 2.92GB βœ… Available 🟒 IMatrix πŸ“¦ No

All Quants

Filename Quant type File Size Status Uses IMatrix Is Split
neo_7b_sft_v0.1.BF16.gguf BF16 15.59GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.FP16.gguf F16 15.59GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q8_0.gguf Q8_0 8.28GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q6_K.gguf Q6_K 6.40GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q5_K.gguf Q5_K 5.54GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q5_K_S.gguf Q5_K_S 5.39GB βœ… Available βšͺ Static πŸ“¦ No
neo_7b_sft_v0.1.Q4_K.gguf Q4_K 4.74GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q4_K_S.gguf Q4_K_S 4.47GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ4_NL.gguf IQ4_NL 4.44GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ4_XS.gguf IQ4_XS 4.20GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q3_K.gguf Q3_K 3.79GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q3_K_L.gguf Q3_K_L 4.11GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q3_K_S.gguf Q3_K_S 3.43GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ3_M.gguf IQ3_M 3.53GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ3_S.gguf IQ3_S 3.43GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ3_XS.gguf IQ3_XS 3.25GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ3_XXS.gguf IQ3_XXS 3.03GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q2_K.gguf Q2_K 2.92GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.Q2_K_S.gguf Q2_K_S 2.71GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ2_M.gguf IQ2_M 2.68GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ2_S.gguf IQ2_S 2.47GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ2_XS.gguf IQ2_XS 2.36GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ2_XXS.gguf IQ2_XXS 2.14GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ1_M.gguf IQ1_M 1.89GB βœ… Available 🟒 IMatrix πŸ“¦ No
neo_7b_sft_v0.1.IQ1_S.gguf IQ1_S 1.73GB βœ… Available 🟒 IMatrix πŸ“¦ No

Downloading using huggingface-cli

If you do not have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Download the specific file you want:

huggingface-cli download legraphista/neo_7b_sft_v0.1-IMat-GGUF --include "neo_7b_sft_v0.1.Q8_0.gguf" --local-dir ./

If the model file is big, it has been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download legraphista/neo_7b_sft_v0.1-IMat-GGUF --include "neo_7b_sft_v0.1.Q8_0/*" --local-dir ./
# see FAQ for merging GGUF's

Inference

Simple chat template

<s>[INST] <<SYS>>
You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe. Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.

If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.
<</SYS>>

{user_prompt} [/INST]{assistant_response}</s><s>[INST] {next_user_prompt} [/INST]

Chat template with system prompt

<s>[INST] {user_prompt} [/INST]{assistant_response}</s><s>[INST] {next_user_prompt} [/INST]

Llama.cpp

llama.cpp/main -m neo_7b_sft_v0.1.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"

FAQ

Why is the IMatrix not applied everywhere?

According to this investigation, it appears that lower quantizations are the only ones that benefit from the imatrix input (as per hellaswag results).

How do I merge a split GGUF?

  1. Make sure you have gguf-split available
  2. Locate your GGUF chunks folder (ex: neo_7b_sft_v0.1.Q8_0)
  3. Run gguf-split --merge neo_7b_sft_v0.1.Q8_0/neo_7b_sft_v0.1.Q8_0-00001-of-XXXXX.gguf neo_7b_sft_v0.1.Q8_0.gguf
    • Make sure to point gguf-split to the first chunk of the split.

Got a suggestion? Ping me @legraphista!

Downloads last month
425
GGUF
Model size
7.79B params
Architecture
llama

1-bit

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Examples
Inference API (serverless) has been turned off for this model.

Model tree for legraphista/neo_7b_sft_v0.1-IMat-GGUF

Quantized
(3)
this model