0.4.0 (2024-09-01)
Features
🚀 examples: integrate Gemma2-2B (#132)
✨ layer_seq: LLM sliding window (#131)
🚀 examples: 3 LLMs examples (#130)
✨ layer_seq: LLM generate (128)
✨ layer_seq: MultiplySeq, SiLU & LLM test (127)
✨ layer_seq: ValueCausalSeq (126)
✨ layer_seq: QueryCausalSeq (125)
✨ layer_seq: RoPESeq (124)
✨ layer_seq: RMSNormSeq (123)
✨ layer_seq: EmbeddingSeq (122)
🪜 feat: LayerCAM2D -> VQGrad2D, LayerCAMSeq -> VQGradSeq (#117)
⚙️ core: GELU vs GELUApprox (113)
🚀 perf: QuerySelf & ValueSelf (112)
🚀 perf: benchmark ViT base model (111)
⚙️ core: initForward,Backward model API (109)
🪜 layer_1d: Dropout1D (#108)
🪜 feat: VQGrad, VQGradSeq (#107)
Bug Fixes
🐛 fix: run on Apple Silicon (110)
Miscellaneous Tasks
📚 docs: LLM doc & split tests (129)
🚀 perf: use half in Metal kernels (121)
🔨 refactor: handle float16 along float on GPU (#120)
🚀 perf: copy & generate weights faster (119)
🚀 perf: Convolution2D (118)