highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-11-11 00:37:42 +01:00
Code Issues Packages Projects Releases Wiki Activity

server: update completion request signature and update token repeat

Browse Source
This commit is contained in:
ParthSareen
2025-08-22 15:40:32 -07:00
parent 51a400ff0f
commit 05cebf1f21
1 changed files with 9 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
llm/server.go
View File

@@ -31,6 +31,7 @@ import (
"github.com/ollama/ollama/envconfig" "github.com/ollama/ollama/envconfig"
"github.com/ollama/ollama/format" "github.com/ollama/ollama/format"
"github.com/ollama/ollama/fs/ggml" "github.com/ollama/ollama/fs/ggml"
"github.com/ollama/ollama/harmony"
"github.com/ollama/ollama/llama" "github.com/ollama/ollama/llama"
"github.com/ollama/ollama/logutil" "github.com/ollama/ollama/logutil"
"github.com/ollama/ollama/ml" "github.com/ollama/ollama/ml"
@@ -1332,6 +1333,8 @@ type CompletionRequest struct {
Options *api.Options Options *api.Options
Grammar string // set before sending the request to the subprocess Grammar string // set before sending the request to the subprocess
FunctionNameMap *harmony.FunctionNameMap
PrefillContent *bool
} }
// DoneReason represents the reason why a completion response is done // DoneReason represents the reason why a completion response is done
@@ -1484,7 +1487,7 @@ func (s *llmServer) Completion(ctx context.Context, req CompletionRequest, fn fu
return fmt.Errorf("error unmarshalling llm prediction response: %v", err) return fmt.Errorf("error unmarshalling llm prediction response: %v", err)
} }
switch { switch {
case strings.TrimSpace(c.Content) == lastToken: case lastToken != "" && (strings.TrimSpace(c.Content) == lastToken || strings.TrimSpace(c.Thinking) == lastToken):
tokenRepeat++ tokenRepeat++
default: default:
lastToken = strings.TrimSpace(c.Content) lastToken = strings.TrimSpace(c.Content)
@@ -1497,14 +1500,14 @@ func (s *llmServer) Completion(ctx context.Context, req CompletionRequest, fn fu
return ctx.Err() return ctx.Err()
} }
if c.Content != "" || c.Thinking != "" || len(c.ToolCalls) > 0 {
fn(c)
}
if c.Done { if c.Done {
fn(c) fn(c)
return nil return nil
} }
if c.Content != "" || c.Thinking != "" || len(c.ToolCalls) > 0 {
fn(c)
}
} }
} }