huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-16 02:38:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix loading the gguf files. (#1913)

Browse Source
This commit is contained in:
Laurent Mazare
2024-03-22 10:28:38 +01:00
committed by GitHub
parent fee33b45c2
commit 32f567bac4
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
candle-transformers/src/models/quantized_llama.rs
View File

@ -362,7 +362,7 @@ impl ModelWeights {
let embedding_length = md_get("llama.embedding_length")?.to_u32()? as usize;
let rope_dim = md_get("llama.rope.dimension_count")?.to_u32()? as usize;
// Strangely this value is generally 1e-6 in GGUF file but used to be 1e-5 by default.
let rms_norm_eps = md_get("llama.attention.layer_norm_rms_epsilon")?.to_f64()?;
let rms_norm_eps = md_get("llama.attention.layer_norm_rms_epsilon")?.to_f32()? as f64;
let rope_freq_base = md_get("llama.rope.freq_base")
.and_then(|m| m.to_f32())