Issues

• Clip IDs and protocol for the Alpamayo 1 open-loop result (937 samples, minADE_6 = 1.22 m)

  NVlabs/alpamayo#121 · Filter by author badkoubeh opened on Oct 20, 2026

• Subject: Inquiry about Flow Matching Export for Alpamayo-R1-10B on TensorRT Edge-LLM

  NVlabs/alpamayo#112 · Filter by author MINTPIPERAS opened on Jul 23, 2026
  

• About the training of RL

  enhancement
  NVlabs/alpamayo#110 · Filter by author DFLyan opened on May 27, 2026
  1

• Does the SFT pipeline support LoRA / PEFT for training on lower-VRAM setups?

  enhancement
  NVlabs/alpamayo#106 · Filter by author minji00116-ctrl opened on May 8, 2026
  3

• Feature suggestion

  enhancement
  NVlabs/alpamayo#70 · Filter by author arthurshelby3210-sudo opened on Apr 20, 2026
  1

• Question about Alpamayo SFT CoC

  NVlabs/alpamayo#68 · Filter by author moonzerokevin opened on Apr 19, 2026
  3

• Evaluating in the closed-loop simulator HUGSIM

  NVlabs/alpamayo#54 · Filter by author TheNeeloy opened on Mar 13, 2026
  4

• Enabling flash_attention_2 causes RuntimeError: cu_seqlens_q must have shape (batch_size + 1)

  NVlabs/alpamayo#52 · Filter by author hyc-hw opened on Mar 2, 2026
  3

• Perception Audit: 312-scene CoC analysis — findings submitted to team

  NVlabs/alpamayo#47 · Filter by author dennisdarrow opened on Feb 12, 2026
  10

• Quick clarification on Alpamayo-R1 training pipeline

  NVlabs/alpamayo#45 · Filter by author mdsgs opened on Feb 11, 2026

• Sensitivity to camera type / numbers - should we expect it to have usable annotation performance on other datasets?

  NVlabs/alpamayo#44 · Filter by author StefanoS90 opened on Feb 6, 2026
  1

• LUT / Color Grading

  NVlabs/alpamayo#41 · Filter by author frrrizzz opened on Jan 30, 2026
  2

• Reasoning-Trajectory Misalignment: "nudge left" reasoning produces right trajectory (reproduced on 4x H100)

  NVlabs/alpamayo#38 · Filter by author dennisdarrow opened on Jan 25, 2026
  6

• to_device() incorrectly casts integer tensors when dtype is already provided

  NVlabs/alpamayo#36 · Filter by author aniekannn opened on Jan 24, 2026

• GT trajectory projection to camera image looks incorrect for PhysicalAI-AV dataset

  NVlabs/alpamayo#34 · Filter by author AutoCR opened on Jan 21, 2026
  1

• How to activate it locally?

  NVlabs/alpamayo#33 · Filter by author haoyu9746-ship-it opened on Jan 20, 2026

• Outline of steps to achieve 99ms latency cited in paper?

  NVlabs/alpamayo#31 · Filter by author yhshin11 opened on Jan 15, 2026
  5

• How to finetune the model to work on a new dataset?

  NVlabs/alpamayo#29 · Filter by author kungfrank opened on Jan 13, 2026
  1

• Question about the VLM backbone

  NVlabs/alpamayo#28 · Filter by author KeqiZeng opened on Jan 13, 2026
  1

• Question: How to couple with Unity or Carla?

  NVlabs/alpamayo#25 · Filter by author M-Colley opened on Jan 12, 2026
  2

• Question about trajectory/action representation

  NVlabs/alpamayo#21 · Filter by author caiocj1 opened on Jan 8, 2026

• Visual concepts recognized by VLM are ignored or forgotten by VLA in safety-critical scenarios

  NVlabs/alpamayo#20 · Filter by author KeqiZeng opened on Jan 8, 2026
  1

• Question about the model's input and output

  NVlabs/alpamayo#14 · Filter by author sonnefred opened on Dec 15, 2025
  8

• [Clarification] Causal relationship score for CoC evaluation procedure

  NVlabs/alpamayo#10 · Filter by author lucamodica opened on Dec 8, 2025

• Questions about visual tokenizer？

  NVlabs/alpamayo#7 · Filter by author LuckySunnye opened on Dec 8, 2025
  5