All Articles · SURL AI News

Transformers Community Reports Memory Growth in Sparse Attention Indexer; Long DeepSeek-V3.2 Prompts May Exhaust GPU Memory

A reproduction on the October 5 development build found that the Lightning Indexer retains full per-attention-head scores, which may sharply increase temporary GPU memory use for long prompts. The report provides an isolated indexer test; effects on the full model and official releases remain unconfirmed.

2026-10-05

GitHub Repo

Dify Community Proposes Scheduling Timezone Fix; New Workflows May Show Local Time but Run on UTC

A Dify 1.17.1 community report on October 5 says schedule nodes whose settings were never edited may fail to save a timezone, causing the backend to schedule them in UTC. The community has proposed a fix, but it has not been merged; teams should also watch for timing changes when republishing existing schedules.

2026-10-05

GitHub Repo

Transformers Community Reports FSDP2 Loading Memory Bug; Multi-GPU Startup May Create Duplicate Full Models

A community report involving Transformers 5.12.0 says that after enabling the CPU memory-efficient loading option, other training processes may still allocate a full model, exhausting host memory. Official documentation says the option should prevent duplicate loading across processes, but the root cause and fix status await upstream confirmation.

2026-10-05

GitHub Repo

Open WebUI community reports document edits out of sync: RAG sees new content, but downloads may still return the old file

A community report dated October 4 about Open WebUI 0.11.4 says document edits may update only the database and vector index, without syncing the original file. This could cause discrepancies in retrieval, downloads, and backups; upstream confirmation is still pending.

2026-10-05

GitHub Repo

Transformers Community Reports RoPE Precision Differences; Converting to bf16 After Loading May Amplify Long-Context Angle Errors

A reproduction case reported on October 4 indicates that converting a model to bf16 after loading may also lower the precision of RoPE frequency buffers, producing results that differ from loading directly in bf16. Maintainers closed a community patch proposal on October 5; the effect on real-model quality and next steps remain to be confirmed.

2026-10-05

GitHub Repo

Transformers Community Proposes MiniMax Precision Fix as bf16 Recurrence Decay May Shift Long-Sequence Decoding

A community reproduction posted on October 4 found that bf16 decay calculations in the development version of Transformers’ MiniMax Lightning Attention may prevent some attention heads from forgetting older information. A fix using fp32 has been proposed but not yet merged; its impact on model quality and performance remains to be validated.

2026-10-05

GitHub Repo

Ollama Community Proposes Request-Parsing Fix After Trailing JSON Noise Is Ignored and Inference Starts

A community reproduction on Ollama 0.35.1 showed that the generate endpoint may accept valid JSON followed by non-JSON text, return success, and run the model. A fix for full-request validation was proposed on October 3 but has not been merged.

2026-10-05

GitHub Repo

Ollama Community Reports MLX Mixed-Quantization Loading Defect: Successful Import May Still Fail at Inference

A community report involving Ollama 0.35.1 suggests that per-layer quantization settings for MLX models may not be applied correctly, causing tensor shape mismatches during loading. The model configuration confirms mixed-precision settings, but the root cause and the scope of impact on other models remain unconfirmed upstream.

2026-10-05

GitHub Repo

PyTorch Community Reports CUDA Interpolation Shortcut Bug That May Skip Resampling for Same-Size Outputs

A reproduction on a PyTorch development build shows that CUDA bicubic interpolation may return the input unchanged when a non-integer scale factor still produces the same output dimensions. A fix has been proposed but not merged; the impact on released versions remains unconfirmed.

2026-10-04

GitHub Repo

Dify Community Reports HTTP Array Serialization Bug That May Silently Alter JSON Requests

Dify’s HTTP Request node was reported to insert object arrays using Python string representation, potentially changing null values and object structure before sending. The community has proposed a JSON serialization fix, but it has not been merged; the affected release range remains unconfirmed.

2026-10-04

GitHub Repo

Transformers Community Proposes Length-Control Fix After Parameter Combination May End Generation Early

Community reproductions show that using a minimum generation length together with an exponential length penalty may produce NaNs, causing generation to end early or sampling to fail. A proposed fix appeared on October 2, but its status in an official release remains unconfirmed.

2026-10-04

GitHub Repo

vLLM Community Reports Batch Invariance Cold-Start Bug That May Consistently Produce Incorrect Tokens on Specific Inference Paths

A reproduction on the vLLM development branch shows that, even with batch invariance enabled, model metadata caching and module import order may still change inference results. The community has proposed a fix, but it has not yet been merged; the impact on released versions and other hardware remains to be confirmed.

2026-09-27