# Getting runtime error when using AutoTrain

**URL:** <https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042>\
**Category:** 🤗AutoTrain\
**Created:** [December 16, 2023, 5:11pm UTC](https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042 "2023-12-16T17:11:34Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![AalianK](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/aaliank/32/22176_2.png) [@AalianK](https://discuss.huggingface.co/u/AalianK)\
**Post date:** [December 16, 2023, 5:11pm UTC](https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042/1 "2023-12-16T17:11:34Z")

</div>

Hello,

I am trying to test out autotrain using a [llama 2 chat](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) and an [alpaca dataset](https://huggingface.co/datasets/tatsu-lab/alpaca) but I have cut it down to 1k rows instead of 52k. I am running on T4 Small

I haven’t changed any configuration options and everything is default.

```JSON
{
  "block_size": 1024,
  "model_max_length": 2048,
  "use_flash_attention_2": false,
  "disable_gradient_checkpointing": false,
  "logging_steps": -1,
  "evaluation_strategy": "epoch",
  "save_total_limit": 1,
  "save_strategy": "epoch",
  "auto_find_batch_size": false,
  "mixed_precision": "fp16",
  "lr": 0.00003,
  "epochs": 3,
  "batch_size": 2,
  "warmup_ratio": 0.1,
  "gradient_accumulation": 1,
  "optimizer": "adamw_torch",
  "scheduler": "linear",
  "weight_decay": 0,
  "max_grad_norm": 1,
  "seed": 42,
  "quantization": "int4",
  "target_modules": "",
  "merge_adapter": false,
  "peft": true,
  "lora_r": 16,
  "lora_alpha": 32,
  "lora_dropout": 0.05
}

```

 ![image](https://us1.discourse-cdn.com/hellohellohello/original/3X/1/8/183b43f3fd6d3cfc53b378b3fe4336206bc647b0.png)

I get the following runtime error

`❌ ERROR | 2023-12-16 17:13:26 | autotrain.trainers.common:wrapper:80 - Expected is_sm80 || is_sm90 to be true, but got false. (Could this error message be improved? If so, please report an enhancement request to PyTorch.)`

I have tried rebuilding with no success

---

<div class="post-metadata">

**Author:** ![latinostats](https://avatars.discourse-cdn.com/v4/letter/l/977dab/32.png) [@latinostats](https://discuss.huggingface.co/u/latinostats)\
**Post date:** [December 22, 2023, 8:57pm UTC](https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042/2 "2023-12-22T20:57:54Z")

</div>

Running into similar problem. Apparently to train LLAMA2 we need GPUs that support sm\_80 or sm\_90. I am having the same error with an RTX 8000 which apparently is sm\_75 capable. Is there a way to know ahead of time which cards are suitable for training a particular LLM?

---

<div class="post-metadata">

**Author:** ![AalianK](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/aaliank/32/22176_2.png) [@AalianK](https://discuss.huggingface.co/u/AalianK)\
**Post date:** [December 22, 2023, 10:06pm UTC](https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042/3 "2023-12-22T22:06:13Z")

</div>

Ran it on A10G and error went away

---

<div class="post-metadata">

**Author:** ![AalianK](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/aaliank/32/22176_2.png) [@AalianK](https://discuss.huggingface.co/u/AalianK)\
**Post date:** [December 22, 2023, 10:09pm UTC](https://discuss.huggingface.co/t/getting-runtime-error-when-using-autotrain/66042/4 "2023-12-22T22:09:49Z")

</div>

I would assume the number of parameters will be a good indicator