fix conversion for f16 or f32 inputs

convert/gemma.go

@@ -1,14 +1,11 @@
 package convert
 
 import (
-	"encoding/binary"
 	"fmt"
 	"io"
 	"log/slog"
-	"os"
 	"strings"
 
-	"github.com/d4l3k/go-bfloat16"
 	"github.com/pdevine/tensor"
 	"github.com/pdevine/tensor/native"
 
@@ -19,49 +16,27 @@ type GemmaModel struct {
 	ModelData
 }
 
-func gemmaLayerHandler(w io.Writer, r safetensorWriterTo, f *os.File) error {
-	slog.Debug(fmt.Sprintf("converting '%s'", r.t.Name))
-
-	data := make([]byte, r.end-r.start)
-	if err := binary.Read(f, r.bo, data); err != nil {
-		return err
-	}
-
-	tDataF32 := bfloat16.DecodeFloat32(data)
-
-	var err error
-	tDataF32, err = addOnes(tDataF32, int(r.t.Shape[0]))
-	if err != nil {
-		return err
-	}
-
-	if err := binary.Write(w, r.bo, tDataF32); err != nil {
-		return err
-	}
-	return nil
-}
-
 func addOnes(data []float32, vectorSize int) ([]float32, error) {
 	n := tensor.New(tensor.WithShape(vectorSize), tensor.WithBacking(data))
 	ones := tensor.Ones(tensor.Float32, vectorSize)
 
-	var err error
+	n, err := n.Add(ones)
-	n, err = n.Add(ones)
 	if err != nil {
-		return []float32{}, err
+		return nil, err
 	}
 
-	newN, err := native.SelectF32(n, 0)
+	ts, err := native.SelectF32(n, 0)
 	if err != nil {
-		return []float32{}, err
+		return nil, err
 	}
 
-	var fullTensor []float32
+	var f32s []float32
-	for _, v := range newN {
+	for _, t := range ts {
-		fullTensor = append(fullTensor, v...)
+		f32s = append(f32s, t...)
 	}
 
-	return fullTensor, nil
+
+	return f32s, nil
 }
 
 func (m *GemmaModel) GetTensors() error {
@@ -74,7 +49,7 @@ func (m *GemmaModel) GetTensors() error {
 	for _, l := range t {
 		if strings.HasSuffix(l.Name, "norm.weight") {
 			wt := l.WriterTo.(safetensorWriterTo)
-			wt.handler = gemmaLayerHandler
+			wt.repacker = m.Repack
 			l.WriterTo = wt
 		}
 		m.Tensors = append(m.Tensors, l)
@@ -92,6 +67,10 @@ func (m *GemmaModel) LoadVocab() error {
 	return nil
 }
 
+func (m *GemmaModel) Repack(_ string, data []float32, shape []uint64) ([]float32, error) {
+	return addOnes(data, int(shape[0]))
+}
+
 func (m *GemmaModel) WriteGGUF(ws io.WriteSeeker) error {
 	kv := llm.KV{
 		"general.architecture":                   "gemma",

convert/llama.go

@@ -1,7 +1,7 @@
 package convert
 
 import (
-	"encoding/binary"
+	"cmp"
 	"errors"
 	"fmt"
 	"io"
@@ -10,10 +10,8 @@ import (
 	"regexp"
 	"strings"
 
-	"github.com/nlpodyssey/gopickle/pytorch"
 	"github.com/pdevine/tensor"
 	"github.com/pdevine/tensor/native"
-	"github.com/x448/float16"
 
 	"github.com/ollama/ollama/llm"
 )
@@ -22,83 +20,6 @@ type LlamaModel struct {
 	ModelData
 }
 
-func llamaTorchLayerHandler(w io.Writer, r torchWriterTo) error {
-
-	var tData []uint16
-	switch r.storage.(type) {
-	case *pytorch.HalfStorage:
-		data := r.storage.(*pytorch.HalfStorage).Data
-		tData = make([]uint16, len(data))
-		for cnt, v := range data {
-			tData[cnt] = uint16(float16.Fromfloat32(v))
-		}
-	case *pytorch.BFloat16Storage:
-		data := r.storage.(*pytorch.BFloat16Storage).Data
-		tData = make([]uint16, len(data))
-
-		for cnt, v := range data {
-			tData[cnt] = uint16(float16.Fromfloat32(v))
-		}
-	default:
-		return fmt.Errorf("unknown storage type for torch")
-	}
-
-	var err error
-	var heads uint32
-	if strings.Contains(r.t.Name, "attn_q") {
-		heads = uint32(r.params.AttentionHeads)
-	} else if strings.Contains(r.t.Name, "attn_k") {
-		heads = uint32(r.params.KeyValHeads)
-		if heads == 0 {
-			heads = uint32(r.params.AttentionHeads)
-		}
-	} else {
-		return fmt.Errorf("unknown layer type")
-	}
-
-	tData, err = llamaRepack(tData, int(heads), r.t.Shape)
-	if err != nil {
-		return err
-	}
-
-	if err = binary.Write(w, r.bo, tData); err != nil {
-		return err
-	}
-	return nil
-}
-
-func llamaRepack(data []uint16, heads int, shape []uint64) ([]uint16, error) {
-	n := tensor.New(tensor.WithShape(int(shape[0]), int(shape[1])), tensor.WithBacking(data))
-	origShape := n.Shape().Clone()
-
-	// reshape the tensor and swap axes 1 and 2 to unpack the layer for gguf
-	if err := n.Reshape(heads, 2, origShape[0]/heads/2, origShape[1]); err != nil {
-		return nil, err
-	}
-
-	if err := n.T(0, 2, 1, 3); err != nil {
-		return nil, err
-	}
-
-	if err := n.Reshape(origShape...); err != nil {
-		return nil, err
-	}
-
-	if err := n.Transpose(); err != nil {
-		return nil, err
-	}
-	newN, err := native.SelectU16(n, 1)
-	if err != nil {
-		return nil, err
-	}
-
-	var fullTensor []uint16
-	for _, v := range newN {
-		fullTensor = append(fullTensor, v...)
-	}
-	return fullTensor, nil
-}
-
 func (m *LlamaModel) GetTensors() error {
 	t, err := m.Format.GetTensors(m.Path, m.Params)
 	if err != nil {
@@ -117,11 +38,11 @@ func (m *LlamaModel) GetTensors() error {
 			switch m.Format.(type) {
 			case *TorchFormat:
 				wt := l.WriterTo.(torchWriterTo)
-				wt.handler = llamaTorchLayerHandler
+				wt.repacker = m.Repack
 				l.WriterTo = wt
 			case *SafetensorFormat:
 				wt := l.WriterTo.(safetensorWriterTo)
-				wt.handler = mistralLayerHandler
+				wt.repacker = m.Repack
 				l.WriterTo = wt
 			}
 		}
@@ -184,3 +105,54 @@ func (m *LlamaModel) WriteGGUF(ws io.WriteSeeker) error {
 
 	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
 }
+
+func (m *LlamaModel) Repack(name string, data []float32, shape []uint64) ([]float32, error) {
+	return llamaRepack(name, m.Params, data, shape)
+}
+
+func llamaRepack(name string, params *Params, data []float32, shape []uint64) ([]float32, error) {
+	var dims []int
+	for _, dim := range shape {
+		if dim != 0 {
+			dims = append(dims, int(dim))
+		}
+	}
+
+	var heads int
+	if strings.HasSuffix(name, "attn_q.weight") {
+		heads = params.AttentionHeads
+	} else if strings.HasSuffix(name, "attn_k.weight") {
+		heads = cmp.Or(params.KeyValHeads, params.AttentionHeads)
+	} else {
+		return nil, fmt.Errorf("unknown tensor name: %s", name)
+	}
+
+	n := tensor.New(tensor.WithShape(dims...), tensor.WithBacking(data))
+	if err := n.Reshape(append([]int{heads, 2, dims[0] / heads / 2}, dims[1:]...)...); err != nil {
+		return nil, err
+	}
+
+	if err := n.T(0, 2, 1, 3); err != nil {
+		return nil, err
+	}
+
+	if err := n.Reshape(dims...); err != nil {
+		return nil, err
+	}
+
+	if err := n.Transpose(); err != nil {
+		return nil, err
+	}
+
+	ts, err := native.SelectF32(n, 1)
+	if err != nil {
+		return nil, err
+	}
+
+	var f32s []float32
+	for _, t := range ts {
+		f32s = append(f32s, t...)
+	}
+
+	return f32s, nil
+}

convert/mistral.go

@@ -1,17 +1,8 @@
 package convert
 
 import (
-	"encoding/binary"
-	"fmt"
 	"io"
-	"os"
 	"regexp"
-	"strings"
-
-	"github.com/d4l3k/go-bfloat16"
-	"github.com/pdevine/tensor"
-	"github.com/pdevine/tensor/native"
-	"github.com/x448/float16"
 
 	"github.com/ollama/ollama/llm"
 )
@@ -20,82 +11,6 @@ type MistralModel struct {
 	ModelData
 }
 
-func mistralLayerHandler(w io.Writer, r safetensorWriterTo, f *os.File) error {
-	layerSize := r.end - r.start
-
-	var err error
-	tData := make([]uint16, layerSize/2)
-	if err = binary.Read(f, r.bo, tData); err != nil {
-		return err
-	}
-
-	var heads uint32
-	if strings.Contains(r.t.Name, "attn_q") {
-		heads = uint32(r.params.AttentionHeads)
-	} else if strings.Contains(r.t.Name, "attn_k") {
-		heads = uint32(r.params.KeyValHeads)
-		if heads == 0 {
-			heads = uint32(r.params.AttentionHeads)
-		}
-	} else {
-		return fmt.Errorf("unknown layer type")
-	}
-
-	tData, err = repack(tData, int(heads), r.t.Shape)
-	if err != nil {
-		return err
-	}
-
-	var buf []byte
-	for _, n := range tData {
-		buf = r.bo.AppendUint16(buf, n)
-	}
-
-	tempBuf := make([]uint16, len(tData))
-	tDataF32 := bfloat16.DecodeFloat32(buf)
-	for cnt, v := range tDataF32 {
-		tDataF16 := float16.Fromfloat32(v)
-		tempBuf[cnt] = uint16(tDataF16)
-	}
-
-	if err = binary.Write(w, r.bo, tempBuf); err != nil {
-		return err
-	}
-	return nil
-}
-
-func repack(data []uint16, heads int, shape []uint64) ([]uint16, error) {
-	n := tensor.New(tensor.WithShape(int(shape[0]), int(shape[1])), tensor.WithBacking(data))
-	origShape := n.Shape().Clone()
-
-	// reshape the tensor and swap axes 1 and 2 to unpack the layer for gguf
-	if err := n.Reshape(heads, 2, origShape[0]/heads/2, origShape[1]); err != nil {
-		return nil, err
-	}
-
-	if err := n.T(0, 2, 1, 3); err != nil {
-		return nil, err
-	}
-
-	if err := n.Reshape(origShape...); err != nil {
-		return nil, err
-	}
-
-	if err := n.Transpose(); err != nil {
-		return nil, err
-	}
-	newN, err := native.SelectU16(n, 1)
-	if err != nil {
-		return nil, err
-	}
-
-	var fullTensor []uint16
-	for _, v := range newN {
-		fullTensor = append(fullTensor, v...)
-	}
-	return fullTensor, nil
-}
-
 func (m *MistralModel) GetTensors() error {
 	t, err := m.Format.GetTensors(m.Path, m.Params)
 	if err != nil {
@@ -112,7 +27,7 @@ func (m *MistralModel) GetTensors() error {
 		matches := re.FindAllStringSubmatch(l.Name, -1)
 		if len(matches) > 0 {
 			wt := l.WriterTo.(safetensorWriterTo)
-			wt.handler = mistralLayerHandler
+			wt.repacker = m.Repack
 			l.WriterTo = wt
 		}
 		m.Tensors = append(m.Tensors, l)
@@ -158,3 +73,7 @@ func (m *MistralModel) WriteGGUF(ws io.WriteSeeker) error {
 
 	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
 }
+
+func (m *MistralModel) Repack(name string, data []float32, shape []uint64) ([]float32, error) {
+	return llamaRepack(name, m.Params, data, shape)
+}

convert/mixtral.go

@@ -27,7 +27,7 @@ func (m *MixtralModel) GetTensors() error {
 		matches := re.FindAllStringSubmatch(l.Name, -1)
 		if len(matches) > 0 {
 			wt := l.WriterTo.(safetensorWriterTo)
-			wt.handler = mistralLayerHandler
+			wt.repacker = m.Repack
 			l.WriterTo = wt
 		}
 		m.Tensors = append(m.Tensors, l)
@@ -81,3 +81,7 @@ func (m *MixtralModel) WriteGGUF(ws io.WriteSeeker) error {
 
 	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
 }
+
+func (m *MixtralModel) Repack(name string, data []float32, shape []uint64) ([]float32, error) {
+	return llamaRepack(name, m.Params, data, shape)
+}

convert/safetensors.go

@@ -27,9 +27,10 @@ type safetensorWriterTo struct {
 	bo     ByteOrder
 
 	filename string
+	dtype    string
 
 	start, end, padding uint64
-	handler             func(w io.Writer, r safetensorWriterTo, f *os.File) error
+	repacker            func(string, []float32, []uint64) ([]float32, error)
 }
 
 type tensorMetaData struct {
@@ -150,6 +151,7 @@ func (m *SafetensorFormat) readTensors(fn string, offset uint64, params *Params)
 			params:   params,
 			bo:       params.ByteOrder,
 			filename: fn,
+			dtype:    data.Type,
 			start:    uint64(data.Offsets[0]),
 			end:      uint64(data.Offsets[1]),
 			padding:  8 + jsonSize,
@@ -235,51 +237,54 @@ func (r safetensorWriterTo) WriteTo(w io.Writer) (n int64, err error) {
 		return 0, err
 	}
 
-	// use the handler if one is present
+	var f32s []float32
-	if r.handler != nil {
+	switch r.dtype {
-		return 0, r.handler(w, r, f)
+	case "F32":
+		f32s = make([]float32, (r.end-r.start)/4)
+		if err = binary.Read(f, r.bo, f32s); err != nil {
+			return 0, err
 		}
-
+	case "F16":
-	remaining := r.end - r.start
+		bts := make([]uint16, (r.end-r.start)/2)
-
+		if err = binary.Read(f, r.bo, bts); err != nil {
-	bufSize := uint64(10240)
-	var finished bool
-	for {
-		data := make([]byte, min(bufSize, remaining))
-
-		b, err := io.ReadFull(f, data)
-		remaining -= uint64(b)
-
-		if err == io.EOF || remaining <= 0 {
-			finished = true
-		} else if err != nil {
 			return 0, err
 		}
 
-		// convert bfloat16 -> ieee float32
+		for _, b := range bts {
-		tDataF32 := bfloat16.DecodeFloat32(data)
+			f32s = append(f32s, float16.Frombits(b).Float32())
+		}
+
+	case "BF16":
+		bts := make([]byte, r.end-r.start)
+		if err = binary.Read(f, r.bo, bts); err != nil {
+			return 0, err
+		}
+
+		f32s = bfloat16.DecodeFloat32(bts)
+	default:
+		return 0, fmt.Errorf("unknown data type: %s", r.dtype)
+	}
+
+	if r.repacker != nil {
+		f32s, err = r.repacker(r.t.Name, f32s, r.t.Shape)
+		if err != nil {
+			return 0, err
+		}
+	}
 
 	switch r.t.Kind {
 	case 0:
-			if err := binary.Write(w, r.bo, tDataF32); err != nil {
+		return 0, binary.Write(w, r.bo, f32s)
-				return 0, err
-			}
 	case 1:
-			// convert float32 -> float16
+		f16s := make([]uint16, len(f32s))
-			tempBuf := make([]uint16, len(data)/2)
+		for i := range f32s {
-			for cnt, v := range tDataF32 {
+			f16s[i] = float16.Fromfloat32(f32s[i]).Bits()
-				tDataF16 := float16.Fromfloat32(v)
-				tempBuf[cnt] = uint16(tDataF16)
 		}
-			if err := binary.Write(w, r.bo, tempBuf); err != nil {
+
-				return 0, err
+		return 0, binary.Write(w, r.bo, f16s)
+	default:
+		return 0, fmt.Errorf("unknown storage type: %d", r.t.Kind)
 	}
-		}
-		if finished {
-			break
-		}
-	}
-	return 0, nil
 }
 
 func (m *SafetensorFormat) GetModelArch(name, dirPath string, params *Params) (ModelArch, error) {

convert/torch.go

@@ -25,7 +25,7 @@ type torchWriterTo struct {
 	bo     ByteOrder
 
 	storage  pytorch.StorageInterface
-	handler func(w io.Writer, r torchWriterTo) error
+	repacker func(string, []float32, []uint64) ([]float32, error)
 }
 
 type TorchFormat struct{}
@@ -230,59 +230,38 @@ func (m *TorchFormat) GetLayerName(n string) (string, error) {
 }
 
 func (r torchWriterTo) WriteTo(w io.Writer) (n int64, err error) {
-	// use the handler if one is present
+	var f32s []float32
-	if r.handler != nil {
+	switch s := r.storage.(type) {
-		return 0, r.handler(w, r)
-	}
-
-	switch storage := r.storage.(type) {
 	case *pytorch.FloatStorage:
-		slog.Warn(fmt.Sprintf("unexpected storage found for layer '%s'; skipping", r.t.Name))
+		f32s = s.Data
-		return 0, nil
 	case *pytorch.HalfStorage:
-		switch r.t.Kind {
+		f32s = s.Data
-		case 0:
-			data := r.storage.(*pytorch.HalfStorage).Data
-			slog.Debug(fmt.Sprintf("%35s F32 (%d)", r.t.Name, len(data)))
-			if err := binary.Write(w, r.bo, data); err != nil {
-				return 0, err
-			}
-		case 1:
-			data := r.storage.(*pytorch.HalfStorage).Data
-			tData := make([]uint16, len(data))
-			for cnt, v := range data {
-				tData[cnt] = uint16(float16.Fromfloat32(v))
-			}
-			slog.Debug(fmt.Sprintf("%35s F16 (%d)", r.t.Name, len(tData)))
-			if err := binary.Write(w, r.bo, tData); err != nil {
-				return 0, err
-			}
-		}
 	case *pytorch.BFloat16Storage:
-		data := r.storage.(*pytorch.BFloat16Storage).Data
+		f32s = s.Data
+	default:
+		return 0, fmt.Errorf("unknown data type: %T", s)
+	}
+
+	if r.repacker != nil {
+		f32s, err = r.repacker(r.t.Name, f32s, r.t.Shape)
+		if err != nil {
+			return 0, err
+		}
+	}
+
 	switch r.t.Kind {
 	case 0:
-			if err = binary.Write(w, r.bo, data); err != nil {
+		return 0, binary.Write(w, r.bo, f32s)
-				return 0, err
-			}
 	case 1:
-			tData := make([]uint16, len(data))
+		f16s := make([]uint16, len(f32s))
-
+		for i := range f32s {
-			for cnt, v := range data {
+			f16s[i] = float16.Fromfloat32(f32s[i]).Bits()
-				tData[cnt] = uint16(float16.Fromfloat32(v))
 		}
 
-			if err = binary.Write(w, r.bo, tData); err != nil {
+		return 0, binary.Write(w, r.bo, f16s)
-				return 0, err
-			}
 	default:
-			return 0, fmt.Errorf("unknown storage kind: %d", r.t.Kind)
+		return 0, fmt.Errorf("unknown storage type: %d", r.t.Kind)
 	}
-	default:
-		return 0, fmt.Errorf("unknown storage type: %T", storage)
-	}
-
-	return 0, nil
 }
 
 func (m *TorchFormat) GetModelArch(name, dirPath string, params *Params) (ModelArch, error) {

go.mod

@@ -4,7 +4,6 @@ go 1.22.0
 
 require (
 	github.com/containerd/console v1.0.3
-	github.com/d4l3k/go-bfloat16 v0.0.0-20211005043715-690c3bdd05f1
 	github.com/emirpasic/gods v1.18.1
 	github.com/gin-gonic/gin v1.10.0
 	github.com/golang/protobuf v1.5.4 // indirect
@@ -18,6 +17,7 @@ require (
 )
 
 require (
+	github.com/d4l3k/go-bfloat16 v0.0.0-20211005043715-690c3bdd05f1
 	github.com/mattn/go-runewidth v0.0.14
 	github.com/nlpodyssey/gopickle v0.3.0
 	github.com/pdevine/tensor v0.0.0-20240510204454-f88f4562727c