diff options
| author | Onur Gündüz <ogunduz@gmail.com> | 2026-05-22 00:25:28 +0200 |
|---|---|---|
| committer | Onur Gündüz <ogunduz@gmail.com> | 2026-06-14 08:18:21 +0200 |
| commit | f1afb07b1ab10324cf0e6607c6b1a4fa018fe2e1 (patch) | |
| tree | c46d2e20e19140fe7d0aad33b5e6232abb3db1ea /pplx_embed.h | |
| parent | f62e7176e6321f21c3a34cfe2064b07c47e19fc3 (diff) | |
| download | ffwd-f1afb07b1ab10324cf0e6607c6b1a4fa018fe2e1.tar.gz ffwd-f1afb07b1ab10324cf0e6607c6b1a4fa018fe2e1.zip | |
cpu: support BF16 weights
Diffstat (limited to 'pplx_embed.h')
| -rw-r--r-- | pplx_embed.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/pplx_embed.h b/pplx_embed.h index 6012576..1bac5f6 100644 --- a/pplx_embed.h +++ b/pplx_embed.h @@ -55,7 +55,7 @@ typedef struct { /* * Load immutable model data from a directory containing: * config.json - model hyperparameters - * model.safetensors (or sharded) - F32 weights + * model.safetensors (or sharded) - F32 or BF16 weights * vocab.json + merges.txt - BPE tokenizer * * Reads config.json to determine model size (0.6B vs 4B). |
