starred/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-04-21 16:25:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

merge ggml file decoding

Browse Source
This commit is contained in:
Michael Yang
2024-12-03 11:17:45 -08:00
parent 2c5fb24855
commit b7943d941d
30 changed files with 429 additions and 1733 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
llm/memory_test.go
View File

@@ -11,6 +11,7 @@ import (
"github.com/ollama/ollama/api"
"github.com/ollama/ollama/discover"
"github.com/ollama/ollama/fs/ggml"
)
func TestEstimateGPULayers(t *testing.T) {
@@ -23,7 +24,7 @@ func TestEstimateGPULayers(t *testing.T) {
defer f.Close()
inputLayerCount := 5
tensors := []Tensor{
tensors := []ggml.Tensor{
{Name: "blk.0.attn.weight", Kind: uint32(0), Offset: uint64(0), Shape: []uint64{1, 1, 1, 1}, WriterTo: bytes.NewReader(make([]byte, 32))},
{Name: "blk.1.attn.weight", Kind: uint32(0), Offset: uint64(0), Shape: []uint64{1, 1, 1, 1}, WriterTo: bytes.NewReader(make([]byte, 32))},
{Name: "blk.2.attn.weight", Kind: uint32(0), Offset: uint64(0), Shape: []uint64{1, 1, 1, 1}, WriterTo: bytes.NewReader(make([]byte, 32))},
@@ -32,7 +33,7 @@ func TestEstimateGPULayers(t *testing.T) {
{Name: "output.weight", Kind: uint32(0), Offset: uint64(0), Shape: []uint64{1, 1, 1, 1}, WriterTo: bytes.NewReader(make([]byte, 32))},
}
assert.Len(t, tensors, inputLayerCount+1)
err = WriteGGUF(f, KV{
err = ggml.WriteGGUF(f, ggml.KV{
"general.architecture": "llama",
"llama.context_length": uint32(32),
"llama.embedding_length": uint32(4096),