Skip to content

Commit

Permalink
[gguf] Add descriptions to quantization types (#615)
Browse files Browse the repository at this point in the history 
I have not found a single place where all different data/quant types of
gguf is documented. Therefore, creating this description object that
would be useful to the community for understanding different data/quant
types.

Afterwards, I plan to make the description available at: 
1. [hf.co/docs/hub/gguf](https://huggingface.co/docs/hub/gguf)
2. [GGUF tensor inspector
](https://twitter.com/mishig25/status/1769731425949884594/video/1)
3. More importantly, community can have a source of information that can
be used in their projects

---------

Co-authored-by: Younes Belkada <[email protected]>
Co-authored-by: FL33TW00D <[email protected]>
  • Loading branch information
@younesbelkada @FL33TW00D
3 people authored Apr 10, 2024
1 parent c87b481 commit 0ed8d60
Show file tree
Hide file tree
Showing 2 changed files with 99 additions and 0 deletions.
1 change: 1 addition & 0 deletions packages/gguf/src/gguf.ts
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ import { GGUFValueType } from "./types";

export type { MetadataBaseValue, MetadataValue, Version, GGUFMetadata, GGUFTensorInfo, GGUFParseOutput } from "./types";
export { GGUFValueType, GGMLQuantizationType } from "./types";
export { QUANT_DESCRIPTIONS } from "./quant-descriptions";

export const RE_GGUF_FILE = /\.gguf$/;
export const RE_GGUF_SHARD_FILE = /-(\d{5})-of-(\d{5})\.gguf$/;
Expand Down
98 changes: 98 additions & 0 deletions packages/gguf/src/quant-descriptions.ts
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,98 @@
import { GGMLQuantizationType } from "./types";

export const QUANT_DESCRIPTIONS: Record<GGMLQuantizationType, { txt: string; src_url?: string }> = {
[GGMLQuantizationType.F32]: {
txt: "32-bit standard IEEE 754 single-precision floating-point number.",
src_url: "https://en.wikipedia.org/wiki/Single-precision_floating-point_format",
},
[GGMLQuantizationType.F16]: {
txt: "16-bit standard IEEE 754 half-precision floating-point number.",
src_url: "https://en.wikipedia.org/wiki/Half-precision_floating-point_format",
},
[GGMLQuantizationType.Q4_0]: {
txt: "4-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557654249",
},
[GGMLQuantizationType.Q4_1]: {
txt: "4-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale + block_minimum. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557682290",
},
[GGMLQuantizationType.Q5_0]: {
txt: "5-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557654249",
},
[GGMLQuantizationType.Q5_1]: {
txt: "5-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale + block_minimum. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557682290",
},
[GGMLQuantizationType.Q8_0]: {
txt: "8-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557654249",
},
[GGMLQuantizationType.Q8_1]: {
txt: "8-bit round-to-nearest quantization (q). Each block has 32 weights. Weight formula: w = q * block_scale + block_minimum. Legacy quantization method (not used widely as of today)",
src_url: "https://github.com/huggingface/huggingface.js/pull/615#discussion_r1557682290",
},
[GGMLQuantizationType.Q2_K]: {
txt: `2-bit quantization (q). Super-blocks with 16 blocks, each block has 16 weight. Weight formula: w = q * block_scale(4-bit) + block_min(4-bit), resulting in 2.5625 bits-per-weight.`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.Q3_K]: {
txt: `3-bit quantization (q). Super-blocks with 16 blocks, each block has 16 weights. Weight formula: w = q * block_scale(6-bit), resulting. 3.4375 bits-per-weight`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.Q4_K]: {
txt: `4-bit quantization (q). Super-blocks with 8 blocks, each block has 32 weights. Weight formula: w = q * block_scale(6-bit) + block_min(6-bit), resulting in 4.5 bits-per-weight.`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.Q5_K]: {
txt: `5-bit quantization (q). Super-blocks with 8 blocks, each block has 32 weights. Weight formula: w = q * block_scale(6-bit) + block_min(6-bit), resulting in 5.5 bits-per-weight.`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.Q6_K]: {
txt: `6-bit quantization (q). Super-blocks with 16 blocks, each block has 16 weights. Weight formula: w = q * block_scale(8-bit), resulting in 6.5625 bits-per-weight.`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.Q8_K]: {
txt: `8-bit quantization (q). Each block has 256 weights. Only used for quantizing intermediate results. All 2-6 bit dot products are implemented for this quantization type. Weight formula: w = q * block_scale.`,
src_url: "https://github.com/ggerganov/llama.cpp/pull/1684#issue-1739619305",
},
[GGMLQuantizationType.IQ2_XXS]: {
txt: "2-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 2.06 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ2_XS]: {
txt: "2-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 2.31 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ3_XXS]: {
txt: "3-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 3.06 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ1_S]: {
txt: "1-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 1.56 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ4_NL]: {
txt: "4-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix",
},
[GGMLQuantizationType.IQ3_S]: {
txt: "3-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 3.44 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ2_S]: {
txt: "2-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 2.5 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
[GGMLQuantizationType.IQ4_XS]: {
txt: "4-bit quantization (q). Super-blocks with 256 weights. Weight w is obtained using super_block_scale & importance matrix, resulting in 4.25 bits-per-weight.",
src_url:
"https://huggingface.co/CISCai/OpenCodeInterpreter-DS-6.7B-SOTA-GGUF/blob/main/README.md?code=true#L59-L70",
},
};

0 comments on commit 0ed8d60

Please sign in to comment.