usage add prompt_tokens

src/models/deepseek_ocr/generate.rs

@@ -102,6 +102,7 @@ impl GenerateModel for DeepseekOCRGenerateModel {
         let mut images_spatial_crop_t = Some(&images_spatial_crop_t);
         let mut seqlen_offset = 0;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut generate = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(1024);
         for _ in 0..sample_len {
@@ -130,7 +131,8 @@ impl GenerateModel for DeepseekOCRGenerateModel {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.deepseekocr_model.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/fun_asr_nano/generate.rs

@@ -114,6 +114,7 @@ impl GenerateModel for FunAsrNanoGenerateModel {
         let mut speech = Some(speech.to_dtype(self.dtype)?);
         let mut fbank_mask = Some(&fbank_mask);
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut generate = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(1024);
@@ -139,7 +140,8 @@ impl GenerateModel for FunAsrNanoGenerateModel {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.fun_asr_nano.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/glm_asr_nano/generate.rs

@@ -77,6 +77,7 @@ impl<'a> GenerateModel for GlmAsrNanoGenerateModel<'a> {
         let mut input_features = Some(input_features.to_dtype(self.dtype)?);
         let mut audio_token_lengths = Some(audio_token_lengths);
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut generate: Vec<u32> = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(1024);
@@ -105,7 +106,8 @@ impl<'a> GenerateModel for GlmAsrNanoGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.glm_asr_nano.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/hunyuan_ocr/generate.rs

@@ -93,6 +93,7 @@ impl<'a> GenerateModel for HunyuanOCRGenerateModel<'a> {
         let mut pixel_values = data.pixel_values;
         let mut image_grid_thw = data.image_grid_thw;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut generate: Vec<u32> = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(1024);
@@ -122,7 +123,8 @@ impl<'a> GenerateModel for HunyuanOCRGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.hunyuan_vl.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/minicpm4/generate.rs

@@ -63,6 +63,7 @@ impl<'a> GenerateModel for MiniCPMGenerateModel<'a> {
         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)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut generate = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(2048);
@@ -81,7 +82,8 @@ impl<'a> GenerateModel for MiniCPMGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.minicpm.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
     fn generate_stream(

src/models/paddleocr_vl/generate.rs

@@ -71,6 +71,7 @@ impl<'a> GenerateModel for PaddleOCRVLGenerateModel<'a> {
             self.pre_processor.process_info(&mes, &mes_render)?;
         let mut input_ids = self.tokenizer.text_encode(replace_text, &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let image_mask = get_equal_mask(&input_ids, self.cfg.image_token_id)?;
 
@@ -107,7 +108,8 @@ impl<'a> GenerateModel for PaddleOCRVLGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.paddleocr_vl.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/qwen2_5vl/generate.rs

@@ -75,6 +75,7 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> {
             .tokenizer
             .text_encode(input.replace_text.clone(), &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut pixel_values = input.pixel_values.as_ref();
         let image_grid_thw = input.image_grid_thw.as_ref();
@@ -121,7 +122,8 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.qwen2_5_vl.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/qwen3/generate.rs

@@ -11,8 +11,8 @@ use crate::models::qwen3::config::{Qwen3Config, Qwen3GenerationConfig};
 use crate::models::qwen3::model::Qwen3Model;
 // use crate::models::GenerateStream;
 use crate::utils::{
-    build_completion_chunk_response, build_completion_response, find_type_files, get_device,
-    get_dtype, get_logit_processor,
+    build_completion_chunk_response, build_completion_response, extract_metadata_value,
+    find_type_files, get_device, get_dtype, get_logit_processor,
 };
 use crate::{chat_template::ChatTemplate, models::GenerateModel, tokenizer::TokenizerModel};
 
@@ -73,9 +73,14 @@ impl<'a> GenerateModel for Qwen3GenerateModel<'a> {
         };
         let mut logit_processor =
             get_logit_processor(Some(temperature), Some(top_p), Some(top_k), seed);
-        let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let enable_thinking = extract_metadata_value::<bool>(&mes.metadata, "enable_thinking");
+        // let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let mes_render = self
+            .chat_template
+            .apply_chat_temp_think(&mes, enable_thinking)?;
         let mut input_ids = self.tokenizer.text_encode(mes_render, &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut generate = Vec::new();
         let sample_len = mes.max_tokens.unwrap_or(2048);
@@ -94,7 +99,8 @@ impl<'a> GenerateModel for Qwen3GenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.qwen3.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
     fn generate_stream(
@@ -123,7 +129,11 @@ impl<'a> GenerateModel for Qwen3GenerateModel<'a> {
         };
         let mut logit_processor =
             get_logit_processor(Some(temperature), Some(top_p), Some(top_k), seed);
-        let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let enable_thinking = extract_metadata_value::<bool>(&mes.metadata, "enable_thinking");
+        // let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let mes_render = self
+            .chat_template
+            .apply_chat_temp_think(&mes, enable_thinking)?;
         let mut input_ids = self.tokenizer.text_encode(mes_render, &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
         let mut seqlen_offset = 0;

src/models/qwen3_5/generate.rs

@@ -74,6 +74,7 @@ impl<'a> GenerateModel for Qwen3_5GenerateModel<'a> {
             .tokenizer
             .text_encode(input.replace_text.clone(), &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut pixel_values = input.pixel_values.as_ref();
         let image_grid_thw = input.image_grid_thw.as_ref();
@@ -102,10 +103,15 @@ impl<'a> GenerateModel for Qwen3_5GenerateModel<'a> {
             pixel_values = None;
             pixel_values_video = None;
         }
-        let num_token = generate.len() as u32;
+        let completion_tokens = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.qwen3_5.clear_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response = build_completion_response(
+            res,
+            &self.model_name,
+            Some(completion_tokens),
+            Some(prompt_tokens),
+        );
         Ok(response)
     }
 

src/models/qwen3_asr/generate.rs

@@ -90,10 +90,12 @@ impl<'a> GenerateModel for Qwen3AsrGenerateModel<'a> {
             .process_info(&mes, &render_text, &self.tokenizer)?;
         let sample_len = mes.max_tokens.unwrap_or(1024);
         let mut generate = Vec::new();
+        let mut prompt_tokens = 0u32;
         for data in audio_datas.iter() {
             let mut input_ids = data.input_ids.clone();
             let mut input_features = Some(data.input_features.clone().to_dtype(self.dtype)?);
             let mut seq_len = input_ids.dim(1)?;
+            prompt_tokens += seq_len as u32;
             let mut seqlen_offset = 0;
             for _ in 0..sample_len {
                 let logits =
@@ -114,7 +116,8 @@ impl<'a> GenerateModel for Qwen3AsrGenerateModel<'a> {
         }
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 

src/models/qwen3vl/generate.rs

@@ -16,8 +16,8 @@ use crate::{
     },
     tokenizer::TokenizerModel,
     utils::{
-        build_completion_chunk_response, build_completion_response, find_type_files, get_device,
-        get_dtype, get_logit_processor,
+        build_completion_chunk_response, build_completion_response, extract_metadata_value,
+        find_type_files, get_device, get_dtype, get_logit_processor,
     },
 };
 
@@ -80,12 +80,17 @@ impl<'a> GenerateModel for Qwen3VLGenerateModel<'a> {
         };
         let mut logit_processor =
             get_logit_processor(Some(temperature), Some(top_p), Some(top_k), seed);
-        let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let enable_thinking = extract_metadata_value::<bool>(&mes.metadata, "enable_thinking");
+        // let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let mes_render = self
+            .chat_template
+            .apply_chat_temp_think(&mes, enable_thinking)?;
         let input = self.pre_processor.process_info(&mes, &mes_render)?;
         let mut input_ids = self
             .tokenizer
             .text_encode(input.replace_text.clone(), &self.device)?;
         let mut seq_len = input_ids.dim(1)?;
+        let prompt_tokens = seq_len as u32;
         let mut seqlen_offset = 0;
         let mut pixel_values = input.pixel_values.as_ref();
         let image_grid_thw = input.image_grid_thw.as_ref();
@@ -120,7 +125,8 @@ impl<'a> GenerateModel for Qwen3VLGenerateModel<'a> {
         let num_token = generate.len() as u32;
         let res = self.tokenizer.token_decode(generate)?;
         self.qwen3_vl.clear_kv_cache();
-        let response = build_completion_response(res, &self.model_name, Some(num_token));
+        let response =
+            build_completion_response(res, &self.model_name, Some(num_token), Some(prompt_tokens));
         Ok(response)
     }
 
@@ -150,7 +156,11 @@ impl<'a> GenerateModel for Qwen3VLGenerateModel<'a> {
         };
         let mut logit_processor =
             get_logit_processor(Some(temperature), Some(top_p), Some(top_k), seed);
-        let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let enable_thinking = extract_metadata_value::<bool>(&mes.metadata, "enable_thinking");
+        // let mes_render = self.chat_template.apply_chat_template(&mes)?;
+        let mes_render = self
+            .chat_template
+            .apply_chat_temp_think(&mes, enable_thinking)?;
         let input = self.pre_processor.process_info(&mes, &mes_render)?;
         let mut input_ids = self
             .tokenizer

src/utils/mod.rs

@@ -479,16 +479,22 @@ pub fn build_audio_completion_response(
 pub fn build_completion_response(
     res: String,
     model_name: &str,
-    num_tokens: Option<u32>,
+    completion_tokens: Option<u32>,
+    prompt_tokens: Option<u32>,
 ) -> ChatCompletionResponse {
     let id = uuid::Uuid::new_v4().to_string();
-    let usage = num_tokens.map(|num| Usage {
-        prompt_tokens: None,
-        completion_tokens: None,
-        total_tokens: num,
+    let usage = if prompt_tokens.is_none() && completion_tokens.is_none() {
+        None
+    } else {
+        Some(Usage {
+            prompt_tokens,
+            completion_tokens,
+            total_tokens: prompt_tokens.unwrap_or(0) + completion_tokens.unwrap_or(0),
             prompt_tokens_details: None,
             completion_tokens_details: None,
-    });
+        })
+    };
+
     let mut response = ChatCompletionResponse {
         id: Some(id),
         choices: vec![],