Problemsby IbisE2B users keep filing bugs
Other2 posts from 2 people2d active+11 post in the last 7 days, 0 the 7 days before (steady)Unverified
Posts per day
The posts behind it
2, newest firstmacOS Apple M4 `failed to prepare readout_recipe` for MiniCPM5 1B (Q8)I just decided to give Magnitude a try, but it doesn't seem to work with MiniCPM5 1B (Q8). The Discover pane said that the best "Balanced" model for my machine is MiniCPM5 1B (Q8), so I downloaded it. I then tried to load it, but the UI in the "My models" pane said: > **This...magnitudedev/magnitude · ForceBru1 repliesOpen ↗ggml-cpu: ggml_vec_dot_bf16 has no NEON implementation - BF16 weights fall back to scalar loop on ARMggml_vec_dot_bf16 (ggml/src/ggml-cpu/vec.cpp) has vectorized branches for AVX512BF16 / AVX512F / AVX2 / RISC-V zvfbfwma / POWER9, but no ARM/NEON branch. On ARMv8.2 (Cortex-A76/A55, __ARM_FEATURE_BF16 absent - that is ARMv8.6+), execution falls to the scalar tail: for (; i < n;...ggml-org/llama.cpp · mrBenBernOpen ↗
Companies and products named
About this problem
Evidence
2 posts from 2 people in 2 places, about 1 a week over 16 days. Mostly on ggml-org/llama.cpp, magnitudedev/magnitude. Tools named alongside: llama.cpp.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts about E2B that state a pain but match no known issue yet. First post Oct 3, 2026, latest Oct 18, 2026. Unverified: fewer than 3 posts, or one person or place only. Method
History
fading to steady
new to fading
E2B users keep filing bugs (1 post)