Repository Issue Activity (beta)

#599 Does BitNet-embedding-0.6B / BitNet-embedding-270M support GPU inference?

#560 Warning: `huggingface-cli` is deprecated and no longer works. Use `hf` instead.

#601 i2_s SIGSEGVs at n_ubatch >= 32: BLAS backend dequantises by a row stride 4x the real packed row

#611 setup_env.py on MacBook M2 failed

#600 BitNet-b1.58-2B-4T produces garbage output on ARM64/NEON (no AVX2) — scalar fallback uses wrong I2_S unpacking scheme

#585 i2_s ARM CPU kernel uses wrong weight-group layout (64/16 vs 128/32) → garbage output on all ARM CPU inference

#610 ARM i2_s (1.58-bit) CPU path produces constant output; fix in pinned llama.cpp submodule

#608 Current bitnet-b1.58-2B-4T checkpoints contain all-zero MLP weight tensors (model generates garbage)

#609 Model Hallucination Occurs again and again.

#588 [Bug]: BitNet FFN uses SILU instead of ReLU² — perplexity 99.8 vs 17.1 on every CPU backend

#602 bitnet-b1.58-2B-4T runs FFN with SiLU instead of relu2 - wrong logits on every backend

#598 I2_S produces incorrect output on aarch64

#595 [Windows] Linker error "undefined symbol: quantize_i2_s" when building with -q i2_s or -q tl2 on Windows 11

#594 Add REFUTE scientific critique + calibration benchmark

#284 Repeated tokens generated from 'generate.py' running on GPU

#569 AVX-512 instructions enable denser packing and more efficient lookups.

#531 Feature Request: Official Discord server for BitNet / bitnet.cpp community

#547 [Bug]: i2_s quantization produces garbage output on x86 CPUs without AVX2 (Missing fallback)

#336 RISC-V Support

#578 deosnt compile and run out of the box

#513 Windows Compilation Fails: CMake errors in setup_env and missing `const` in ggml-bitnet-mad

#407 Compiler error

#561 Here are the current steps for builing BitNet.cpp for Raspberry Pi 5

#378 CMake build fails due to missing `bitnet-lut-kernels.h`

#557 Debian package with systemd socket activation (bitnet-server)