TinyLLaVA/TinyLLaVA_Factory

Issues

• Unable to use GPUs 4,5,6,7 for training

  #112 opened 16 days ago by shashwat14
  3

• Gemma template for integrating instruct tuned model

  #96 opened 17 days ago by Jayantverma2
  1

• About "conv_version" and PretrainTemplate .

  #101 opened 17 days ago by NyKxo1
  1

• qformer

  #110 opened 17 days ago by PangziZhang523
  1

• About the Vision Tower.

  #111 opened 21 days ago by Sootung
  1

• Finetuning get: ValueError: not enough values to unpack (expected 2, got 0) in get_modality_length_grouped_indices

  #109 opened 21 days ago by sc268
  2

• Continue fine-tuning from previously trained model: error on inference

  #104 opened a month ago by simoneriggi
  8

• About share recipe. Should change "tune_type_vision_tower" to set to "partially-tune"?

  #102 opened a month ago by NyKxo1
  1

• Does the Phi template work for Phi3？

  #103 opened a month ago by NyKxo1
  1

• Add '--raise_error_at_min_scale False' flag in pretrain and finetune scripts to avoid minimum loss scale crash caused by bad batch.

  #107 opened a month ago by RobotiX101
  0

• 在使用hf上下载TinyLLaVA-Phi-2-SigLIP-3.1B模型时 ，参数设置。

  #105 opened a month ago by eva10084
  0

• replace llm

  #95 opened 2 months ago by LilDevsy0117
  1

• Can you reproduce the Llava results in this project?

  #97 opened 2 months ago by Yangr116
  2

• TypeError: __init__() takes 1 positional argument but 2 were given

  #100 opened a month ago by RobotiX101
  0

• template differences？

  #77 opened 3 months ago by TuuSiwei
  2

• Exception: Current Loss Scale Already at Minimum While Fine-Tuning with 8 A800-80G GPUs

  #99 opened 2 months ago by codefanw
  2

• Config for TinyLLaVA-OpenELM-270M-SigLIP-0.55B

  #70 opened 3 months ago by eternalding
  4

• distributed computing

  #93 opened 2 months ago by 1764758458
  1

• mof_mlp error

  #94 opened 2 months ago by Daming-W
  1

• require_grad

  #81 opened 3 months ago by 1764758458
  6

• Results are not reproducible on Qwen1.5-1.8B

  #92 opened 2 months ago by Fantasy1120
  0

• dataset seems imcomplete

  #91 opened 2 months ago by BroJunn
  1

• Why `group_by_modality` is set to False on the `custom_finetune.sh` script?

  #89 opened 3 months ago by ggcr
  3

• Missing key "lm_head.weight" in GemmaForCausalLM when loading lora finetuned TinyLLaVA-Gemma-SigLIP-2.4B

  #88 opened 3 months ago by Yuki-Kokomi
  5

• really_cleaned_share-captioner_coco_lcs_sam_1246k_1107.json not found

  #86 opened 3 months ago by Wangman1
  2

• Create a tinyllava with qwen2-0.5B-instruct

  #87 opened 3 months ago by bil-ash
  0

• demo link is down.

  #85 opened 3 months ago by cooleel
  1

• Questions about LLM templates

  #83 opened 3 months ago by ShawnAn-WHU
  1

• Reproduced results

  #82 opened 3 months ago by Fantasy1120
  3

• How to inference multiple images from custom finetuned data?

  #79 opened 3 months ago by hengtseChou
  2

• --fp16 True question

  #78 opened 3 months ago by Liavan0122
  6

• share training recipe

  #80 opened 3 months ago by Fantasy1120
  1

• Data about evaluation

  #75 opened 3 months ago by Fantasy1120
  2

• Implementing RLHF trainer ?

  #74 opened 3 months ago by R3xpook
  3

• Question about 'fix everything'

  #71 opened 3 months ago by Zeqing-Wang
  3

• Does Tiny LLaVa support multi-turn and multi-images conversation?

  #73 opened 3 months ago by Liuziyu77
  1

• Mistake in eval_textvqa

  #72 opened 3 months ago by hedes1992
  3

• What does "share" mean in the last line of the model performance table.

  #69 opened 3 months ago by swordlidev
  1

• Replace vision tower with DINOv2

  #68 opened 4 months ago by Daming-W
  1

• consuming time for each VLM pretrain/finetune

  #67 opened 4 months ago by vision-mini
  2

• Fine-tuning TinyLLaVA-Phi-2-SigLIP-3.1B

  #62 opened 4 months ago by saadi297
  8

• ModuleNotFoundError: No module named 'tinyllava.model.builder'

  #63 opened 4 months ago by dkjung
  3

• MM-Vet download link broken

  #66 opened 4 months ago by ryan-caesar-ramos
  2

• Error when using the OpenELM as the LLM backbone

  #64 opened 4 months ago by Zeqing-Wang
  5

• Is share_textvqa is the same as textvqa in llava dataset?

  #65 opened 4 months ago by cooleel
  1

• Request for Weights of apple/OpenELM-450M-Instruct Model

  #61 opened 4 months ago by rotem154154
  3

• add phi3

  #59 opened 4 months ago by Jayantverma2
  0

• question about custom fine-tuning

  #58 opened 4 months ago by josephtey
  0

• Missing `llavabench.sh`

  #57 opened 4 months ago by cooleel
  0

• Lora finetune results drop dramatically

  #56 opened 4 months ago by YFCYFC
  2