mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-24 07:28:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'ollama:main' into main

Browse Source
This commit is contained in:
likelovewant
2024-05-26 11:57:21 +08:00
committed by GitHub
parent 73c49d57e8 0165ba1651
commit 2a80d6f743
16 changed files with 85 additions and 55 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llm/ggml.go
View File

@@ -125,9 +125,9 @@ type Tensor struct {
func (t Tensor) blockSize() uint64 {
switch t.Kind {
case 0, 1, 24, 25, 26, 27, 28, 31: // F32, F16, I8, I16, I32, I64, F64, BF16
case 0, 1, 24, 25, 26, 27, 28, 30: // F32, F16, I8, I16, I32, I64, F64, BF16
return 1
case 2, 3, 8, 9, 20: // Q4_0, Q4_1, Q8_0, Q8_1, IQ4_NL
case 2, 3, 4, 5, 6, 7, 8, 9, 20: // Q4_0, Q4_1, Q5_0, Q5_1, Q8_0, Q8_1, IQ4_NL
return 32
default: // All others
return 256

2
llm/memory.go
View File

@@ -7,7 +7,7 @@ import (
"github.com/ollama/ollama/api"
"github.com/ollama/ollama/format"
"github.com/ollama/ollama/gpu"
"github.com/ollama/ollama/server/envconfig"
"github.com/ollama/ollama/envconfig"
)
// This algorithm looks for a complete fit to determine if we need to unload other models

4
llm/server.go
View File

@@ -26,7 +26,7 @@ import (
"github.com/ollama/ollama/api"
"github.com/ollama/ollama/format"
"github.com/ollama/ollama/gpu"
"github.com/ollama/ollama/server/envconfig"
"github.com/ollama/ollama/envconfig"
)
type LlamaServer interface {
@@ -528,7 +528,7 @@ func (s *llmServer) WaitUntilRunning(ctx context.Context) error {
for {
select {
case <-ctx.Done():
slog.Info("context expired before server started")
slog.Warn("client connection closed before server finished loading, aborting load")
return fmt.Errorf("timed out waiting for llama runner to start: %w", ctx.Err())
case err := <-s.done:
msg := ""