huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-20 20:09:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

Documentation Pass for Models (#2617)

Browse Source 
* links in chinese_clip

* links for clip model

* add mod docs for flux and llava

* module doc for MMDIT and MIMI

* add docs for a few more modesl

* mod docs for bert naser and beit

* add module docs for convmixer colpali codegeex and chatglm

* add another series of moddocs

* add  fastvit-llama2_c

* module docs mamba -> mobileone

* module docs from moondream-phi3

* mod docs for quantized and qwen

* update to yi

* fix long names

* Update llama2_c.rs

* Update llama2_c_weights.rs

* Fix the link for mimi + tweaks

---------

Co-authored-by: Laurent Mazare <laurent.mazare@gmail.com>
This commit is contained in:
zachcp
2024-11-15 02:30:15 -05:00
committed by GitHub
parent 0ed24b9852
commit f689ce5d39
94 changed files with 1001 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
candle-transformers/src/models/quantized_rwkv_v6.rs
View File

@ -1,3 +1,21 @@
//! RWKV v6 model implementation with quantization support.
//!
//! RWKV is a linear attention model that combines the efficiency of RNNs
//! with the parallelizable training of Transformers. Version 6 builds on previous
//! versions with further optimizations.
//!
//! Key characteristics:
//! - Linear attention mechanism
//! - Time mixing layers
//! - Channel mixing layers
//! - RMSNorm for normalization
//! - Support for 8-bit quantization
//!
//! References:
//! - [RWKV Architecture](https://github.com/BlinkDL/RWKV-LM)
//! - [RWKV v6 Release](https://huggingface.co/BlinkDL/rwkv-6)
//!
use crate::{
quantized_nn::{layer_norm, linear_no_bias as linear, Embedding, Linear},
quantized_var_builder::VarBuilder,