Problemsby GlyphAll problems
133 problems
Running Perplexity locally hits memory and speed limits
3 posts from 3 people on GitHub issues.
Trouble with PyTorch inputs
Top post: "integer torch.bincount is ~1000x slower when most inputs fall in one bin", one of 3 posts from 3 people on GitHub issues.
Trouble with NVIDIA process
Top post: "`HuggingFaceEmbeddings(multi_process=True)` starts CPU workers on non-CUDA accelerators", one of 3 posts from 3 people on GitHub issues.
DeepSeek rate limits and usage caps block work
3 posts from 3 people on Hacker News, GitHub issues, Reddit.
Trouble with OpenAI raises
Top post: "Whisper generate() raises 'You have modified the pretrained model configuration' since #48282", one of 3 posts from 3 people on GitHub issues.
NVIDIA runs out of context on real work
2 posts from 2 people on GitHub issues. Unverified.
Trouble with PyTorch load
Top post: "Qwen3-VL-Reranker fails to load because Qwen3VLTextConfig has no tie_word_embeddings", one of 2 posts from 2 people on GitHub issues. Unverified.
Serving models on Hugging Face is harder than it should be
2 posts from 2 people on GitHub issues. Unverified.
Trouble with NVIDIA verification
Top post: "Unable to receive SMS verification code for NVIDIA Build API key", one of 2 posts from 2 people on Community forums. Unverified.
NVIDIA vulnerabilities are hard to triage and patch
77 posts from 77 people on Glyph Ibis.
AI-generated code is buggy and hard to review
41 posts from 41 people on Hacker News, Reddit, App Store reviews.
LLM token costs are hard to predict and control
14 posts from 14 people on GitHub issues, Reddit, Hacker News.
Trouble with GitHub Copilot agent
Top post: "Copilot Agent edit review workflow became less intuitive in VS Code 1.143.x", one of 12 posts from 12 people on GitHub issues, Reddit.
NVIDIA upgrades break working setups
11 posts from 11 people on GitHub issues.
Trouble with NVIDIA account
Top post: "NVIDIA Careers Workday Account Locked - Unable to Complete Employee Referral Application", one of 9 posts from 8 people on Community forums.
Fine-tuning Hugging Face is costly and unpredictable
9 posts from 9 people on GitHub issues, Reddit.
Ollama runs out of context on real work
7 posts from 7 people on GitHub issues, Reddit.
Trouble with Ollama empty
Top post: "Ollama photo/label estimation returns empty JSON with reasoning models (gemma4, qwen3.6), request should send", one of 7 posts from 6 people on GitHub issues.
Trouble with PyTorch input
Top post: "Generation crashes with encoder_repetition_penalty when encoder_outputs are provided without input_ids", one of 6 posts from 6 people on GitHub issues.
Serving models on Ollama is harder than it should be
6 posts from 6 people on GitHub issues, Reddit.
Trouble with Hugging Face generate
Top post: "prefill_chunk_size drops image inputs, VLMs generate without the image", one of 6 posts from 5 people on GitHub issues, Community forums.
Trouble with Ollama pull
Top post: "Failed to pull `embeddinggemma-2:740m` on linux", one of 6 posts from 6 people on GitHub issues.
Trouble with llama.cpp compile
Top post: "Compile bug: rocm-6.3.4, no matching function for call to group_mask_warp_inplace", one of 5 posts from 5 people on GitHub issues.
NVIDIA lock-in makes switching costly
5 posts from 5 people on GitHub issues, Hacker News, Bluesky.
Trouble with Claude Code sessions
Top post: "Active sessions just disappear", one of 5 posts from 5 people on GitHub issues, Reddit.
Trouble with Ollama text
Top post: "Ollama Text Classifier modles slower under version 0.40.4 ?", one of 4 posts from 3 people on GitHub issues.
Trouble with Hugging Face logits
Top post: "`DogeCDMoE` router `.view(2, bsz * seq_len, -1)` mixes tokens: with `is_moe=True`, logits depend on future", one of 4 posts from 4 people on GitHub issues.
Trouble with vLLM cache
Top post: "AOT compile cache never loads for Qwen3.5 models: Kernel index 1 not found in id_to_kernel", one of 4 posts from 4 people on GitHub issues.
Trouble with Hugging Face torch
Top post: "import transformers spends 1-3 s in an eager importlib.metadata.packages_distributions() when torch >= 2.14", one of 4 posts from 4 people on GitHub issues.
Fine-tuning Qwen is costly and unpredictable
4 posts from 4 people on GitHub issues, Hacker News, Reddit.
Trouble with vLLM serve
Top post: "qwen4_exp (Qwen3.8-Flash-Next) cannot serve DeepSpec DFlash/DSpark drafters: five blockers, patches available", one of 3 posts from 3 people on GitHub issues.
Ollama GPU access and drivers are painful
3 posts from 3 people on Hacker News, Reddit, GitHub issues.
Trouble with DeepSeek flash
Top post: "Asymmetric P/D Deployment for DeepSeek-V4.1 Flash", one of 3 posts from 3 people on GitHub issues, Community forums.
Fine-tuning PyTorch is costly and unpredictable
3 posts from 3 people on GitHub issues, Reddit.
Trouble with PyTorch drops
Top post: "constant-False cond_fn drops captured-tensor mutations", one of 3 posts from 3 people on GitHub issues.
Trouble with PyTorch torch
Top post: "5.19.0: models cannot be imported on torch 2.5, or on torch 2.6 without an accelerator (get_device_type calls", one of 3 posts from 3 people on GitHub issues.
Trouble with NVIDIA phone
Top post: "Unable to receive OTP for Vietnam (+84) phone verification on build.nvidia.com", one of 3 posts from 3 people on Community forums.
Trouble with Hugging Face size
Top post: "Inkling-Small: width of the routed experts resolves to 3072 instead of 2048 (moe_intermediate_size is never", one of 3 posts from 3 people on GitHub issues.
Trouble with vLLM speculative
Top post: "Support Qwen3.x MTP (Multi-Token Prediction) for speculative decoding", one of 3 posts from 3 people on GitHub issues.