mirror of
https://github.com/dogkeeper886/ollama37.git
synced 2025-12-18 19:56:59 +00:00
ef14fb5b26
This commit represents a complete rework after pulling the latest changes from official ollama/ollama repository and re-applying Tesla K80 compatibility patches. ## Key Changes ### CUDA Compute Capability 3.7 Support (Tesla K80) - Added sm_37 (compute 3.7) to CMAKE_CUDA_ARCHITECTURES in CMakeLists.txt - Updated CMakePresets.json to include compute 3.7 in "CUDA 11" preset - Using 37-virtual (PTX with JIT compilation) for maximum compatibility ### Legacy Toolchain Compatibility - **NVIDIA Driver**: 470.256.02 (last version supporting Kepler/K80) - **CUDA Version**: 11.4.4 (last CUDA 11.x supporting compute 3.7) - **GCC Version**: 10.5.0 (required by CUDA 11.4 host_config.h) ### CPU Architecture Trade-offs Due to GCC 10.5 limitation, sacrificed newer CPU optimizations: - Alderlake CPU variant enabled WITHOUT AVX_VNNI (requires GCC 11+) - Still supports: SSE4.2, AVX, F16C, AVX2, BMI2, FMA - Performance impact: ~3-7% on newer CPUs (acceptable for K80 compatibility) ### Build System Updates - Modified ml/backend/ggml/ggml/src/ggml-cuda/CMakeLists.txt for compute 3.7 - Added -Wno-deprecated-gpu-targets flag to suppress warnings - Updated ml/backend/ggml/ggml/src/CMakeLists.txt for Alderlake without AVX_VNNI ### Upstream Sync Merged latest llama.cpp changes including: - Enhanced KV cache management with ISWA and hybrid memory support - Improved multi-modal support (mtmd framework) - New model architectures (Gemma3, Llama4, Qwen3, etc.) - GPU backend improvements for CUDA, Metal, and ROCm - Updated quantization support and GGUF format handling ### Documentation - Updated CLAUDE.md with comprehensive build instructions - Documented toolchain constraints and CPU architecture trade-offs - Removed outdated CI/CD workflows (tesla-k80-*.yml) - Cleaned up temporary development artifacts ## Rationale This fork maintains Tesla K80 GPU support (compute 3.7) which was dropped in official Ollama due to legacy driver/CUDA requirements. The toolchain constraint creates a deadlock: - K80 → Driver 470 → CUDA 11.4 → GCC 10 → No AVX_VNNI We accept the loss of cutting-edge CPU optimizations to enable running modern LLMs on legacy but still capable Tesla K80 hardware (12GB VRAM per GPU). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
74 lines
2.2 KiB
Go
74 lines
2.2 KiB
Go
package qwen3
|
|
|
|
import (
|
|
"github.com/ollama/ollama/fs"
|
|
"github.com/ollama/ollama/kvcache"
|
|
"github.com/ollama/ollama/ml"
|
|
"github.com/ollama/ollama/ml/nn/pooling"
|
|
"github.com/ollama/ollama/model"
|
|
"github.com/ollama/ollama/model/input"
|
|
)
|
|
|
|
type embedModel struct {
|
|
model.Base
|
|
model.BytePairEncoding
|
|
|
|
*Model
|
|
poolingType pooling.Type
|
|
}
|
|
|
|
func (m *embedModel) Forward(ctx ml.Context, batch input.Batch) (ml.Tensor, error) {
|
|
hiddenStates, err := m.forward(ctx, batch)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
hiddenStates = m.poolingType.Forward(ctx, hiddenStates)
|
|
hiddenStates = hiddenStates.L2Norm(ctx, 1e-12)
|
|
return hiddenStates, nil
|
|
}
|
|
|
|
func newEmbed(c fs.Config) (model.Model, error) {
|
|
layers := make([]Layer, c.Uint("block_count"))
|
|
for i := range layers {
|
|
layers[i].MLP = &dense{}
|
|
}
|
|
m := embedModel{
|
|
BytePairEncoding: model.NewBytePairEncoding(
|
|
&model.Vocabulary{
|
|
Values: c.Strings("tokenizer.ggml.tokens"),
|
|
Types: c.Ints("tokenizer.ggml.token_type"),
|
|
Merges: c.Strings("tokenizer.ggml.merges"),
|
|
AddBOS: c.Bool("tokenizer.ggml.add_bos_token", true),
|
|
BOS: []int32{int32(c.Uint("tokenizer.ggml.bos_token_id"))},
|
|
AddEOS: c.Bool("tokenizer.ggml.add_eos_token", false),
|
|
EOS: append(
|
|
[]int32{int32(c.Uint("tokenizer.ggml.eos_token_id"))},
|
|
c.Ints("tokenizer.ggml.eos_token_ids")...,
|
|
),
|
|
},
|
|
`(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+`,
|
|
),
|
|
Model: &Model{
|
|
Layers: layers,
|
|
Options: &Options{
|
|
hiddenSize: int(c.Uint("embedding_length")),
|
|
numHeads: int(c.Uint("attention.head_count")),
|
|
numKVHeads: int(c.Uint("attention.head_count_kv")),
|
|
keyLength: int(c.Uint("attention.key_length")),
|
|
valueLength: int(c.Uint("attention.value_length")),
|
|
eps: c.Float("attention.layer_norm_rms_epsilon"),
|
|
ropeBase: c.Float("rope.freq_base"),
|
|
ropeScale: c.Float("rope.freq_scale", 1),
|
|
numExperts: int(c.Uint("expert_count")),
|
|
numExpertsUsed: int(c.Uint("expert_used_count")),
|
|
normTopKProb: c.Bool("norm_top_k_prob", true),
|
|
},
|
|
},
|
|
poolingType: pooling.Type(c.Uint("pooling_type")),
|
|
}
|
|
|
|
m.Cache = kvcache.NewCausalCache(m.Shift)
|
|
return &m, nil
|
|
}
|