huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-04-21 23:12:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

Make bf16 default for hpu, fix script (#205)

Browse Source
This commit is contained in:
Abhilash Majumder 2024-08-11 14:18:35 +05:30 committed by GitHub
parent cf2ff5a1dd
commit d403575c43
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
launcher/src/main.rs
View File

@ -182,7 +182,7 @@ struct Args {
speculate: Option<usize>,
/// The dtype to be forced upon the model. This option cannot be used with `--quantize`.
#[clap(long, env, value_enum)]
#[clap(default_value = "bfloat16", long, env, value_enum)]
dtype: Option<Dtype>,
/// Whether you want to execute hub modelling code. Explicitly passing a `revision` is
@ -501,13 +501,12 @@ fn shard_manager(
if let Some(dtype) = dtype {
shard_args.push("--dtype".to_string());
shard_args.push(dtype.to_string())
shard_args.push(dtype.to_string());
}
// Model optional revision
if let Some(revision) = revision {
shard_args.push("--revision".to_string());
shard_args.push(revision)
shard_args.push(revision);
}
let rope = match (rope_scaling, rope_factor) {