Skip to content

FSDP+QLoRA ValueError: Must flatten tensors with uniform dtype (float32 vs bfloat16)

FSDP combined with QLoRA fails to build its flat parameter: 'Must flatten tensors with uniform dtype but got torch.float32 and torch.bfloat16'. QLoRA keeps some parameters in float32 while the base is bfloat16; FSDP requires one dtype per flattened parameter group. Use the maintained FSDP+QLoRA config and consistent dtypes.

Quick answer

FSDP combined with QLoRA fails to build its flat parameter: 'Must flatten tensors with uniform dtype but got torch.

Distributed Training#fsdp#qlora#mixed-precision#dtype#axolotl#flat-param

What this failure is

FSDP+QLoRA ValueError: Must flatten tensors with uniform dtype (float32 vs bfloat16) is a Distributed Training failure seen during ML training runs. FSDP combined with QLoRA fails to build its flat parameter: 'Must flatten tensors with uniform dtype but got torch.float32 and torch.bfloat16'. QLoRA keeps some parameters in float32 while the base is bfloat16; FSDP requires one dtype per flattened parameter group. Use the maintained FSDP+QLoRA config and consistent dtypes. Common tags: Fsdp, Qlora, Mixed Precision, Dtype.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about FSDP+QLoRA ValueError: Must flatten tensors with uniform dtype (float32 vs bfloat16). Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.

training_logs.txt
No log to hand? Try one:

3 free diagnoses/day

Want 14 days on the Scale plan?

Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.

Evaluate one incident

Why it happens (the mechanism)

QLoRA keeps LoRA adapters and some norm/embedding parameters in float32 while the quantized base computes in bfloat16. FSDP flattens a parameter group into a single flat tensor and requires every parameter in that group to share one dtype. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.

What you'll observe

  • FSDP wrapping fails during flat-param construction
  • Mixed float32/bfloat16 parameters in one FSDP group
  • Llama-2 / Mixtral QLoRA-FSDP recipes crash at startup

Common symptoms and what they mean

SymptomWhy it happens
ValueError: Must flatten tensors with uniform dtype but got torch.float32 and torch.bfloat16QLoRA keeps LoRA adapters and some norm/embedding parameters in float32 while the quantized base computes in bfloat16
Error while FSDP flattens a parameter groupFSDP flattens a parameter group into a single flat tensor and requires every parameter in that group to share one dtype
Happens with QLoRA (4-bit base) under FSDPQLoRA keeps LoRA adapters and some norm/embedding parameters in float32 while the quantized base computes in bfloat16

Which systems are affected

  • Axolotl/TRL FSDP + QLoRA fine-tuning
  • FSDP flat-parameter wrapping of mixed-dtype modules
  • Quantized base + float32 LoRA/norm parameters

How to confirm this is the problem

Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.

  • Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
  • Verified signal present: ValueError: Must flatten tensors with uniform dtype but got torch.float32 and torch.bfloat16
  • Verified signal present: Error while FSDP flattens a parameter group
  • Verified signal present: Happens with QLoRA (4-bit base) under FSDP
  • A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

The fix and the prevention pattern

The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.

Sign up free. Unlock the full analysis

No credit card. Daily allowance follows verified trust tier. Instant access.

Diagnose this failure in VS Code

Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.

Install the free VS Code extension

Root cause

  • QLoRA keeps LoRA adapters and some norm/embedding parameters in float32 while the quantized base computes in bfloat16
  • FSDP flattens a parameter group into a single flat tensor and requires every parameter in that group to share one dtype

The fix and how to prevent it

Evaluate Denpex on your own logs

Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.

We send a single-use code tied to that address. Static provider and TLD rules do not reject valid addresses. Account trust determines the benefit after signup.

Don't just read the fix, diagnose your run

The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.