huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-06-19 15:52:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove warmup

Browse Source 
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët 2025-02-01 20:27:31 +00:00
parent 8ed362d03a
commit 104a968d01
No known key found for this signature in database
1 changed files with 0 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
backends/llamacpp/src/backend.rs
View File

@ -238,38 +238,6 @@ impl Llamacpp {
self.batch.n_tokens += 1;
n
}
// useless ?
fn warmup(&self) {
let mut buf: Vec<bindings::llama_token> = Vec::new();
let bos = unsafe {
bindings::llama_vocab_bos(self.vocab)
};
if bos != bindings::LLAMA_TOKEN_NULL {
buf.push(bos);
}
let eos = unsafe {
bindings::llama_vocab_eos(self.vocab)
};
if eos != bindings::LLAMA_TOKEN_NULL {
buf.push(eos);
}
if buf.is_empty() {
warn!("Warmup failed: no bos/eos...");
return;
}
let batch = unsafe {
bindings::llama_batch_get_one(buf.as_ptr() as _, buf.len() as _)
};
if unsafe { bindings::llama_decode(self.ctx, batch) } != 0 {
error!("Warmup failed: llama_decode() returned an error");
}
unsafe {
bindings::llama_kv_cache_clear(self.ctx);
bindings::llama_synchronize(self.ctx);
}
}
}
impl Drop for Llamacpp {
@ -456,8 +424,6 @@ impl LlamacppBackend {
Ok(v) => { let _ = ok_tx.send(Ok(())); v },
Err(e) => { let _ = ok_tx.send(Err(e)); return; },
};
llamacpp.warmup();
let vocab = tokenizer.get_added_vocabulary();
// health() returns true