tckm.yan/aha
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

add Qwen3VL model

Browse Source
This commit is contained in:
jhqxxx
2025-10-26 21:39:23 +08:00
parent cdae7fc45f
commit 7b190e7d54
22 changed files with 2591 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models/minicpm4/generate.rs
+2 -2
View File
@@ -53,7 +53,7 @@ impl<'a> MiniCPMGenerateModel<'a> {
impl<'a> GenerateModel for MiniCPMGenerateModel<'a> {
fn generate(&mut self, mes: ChatCompletionParameters) -> Result<ChatCompletionResponse> {
let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p);
let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p, None);
let mes_render = self.chat_template.apply_chat_template(&mes)?;
let mut input_ids = self.tokenizer.text_encode(mes_render, &self.device)?;
let mut seq_len = input_ids.dim(1)?;
@@ -81,7 +81,7 @@ impl<'a> GenerateModel for MiniCPMGenerateModel<'a> {
&mut self,
mes: ChatCompletionParameters,
) -> Result<impl Stream<Item = Result<ChatCompletionChunkResponse, anyhow::Error>>> {
let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p);
let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p, None);
let mes_render = self.chat_template.apply_chat_template(&mes)?;
let mut input_ids = self.tokenizer.text_encode(mes_render, &self.device)?;
let mut seq_len = input_ids.dim(1)?;