Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / TinyLLaVA/TinyLLaVA_Factory issues and pull requests
#115 - Export To Common Formats
Issue -
State: open - Opened by TalonBvV 7 days ago
- 1 comment
#114 - Model type `tinyllava` not recognized by Transformers
Issue -
State: closed - Opened by AaronChuh 10 days ago
#113 - surpport on mobile device
Issue -
State: open - Opened by perp 14 days ago
#112 - Unable to use GPUs 4,5,6,7 for training
Issue -
State: open - Opened by shashwat14 23 days ago
- 3 comments
#111 - About the Vision Tower.
Issue -
State: open - Opened by Sootung 28 days ago
- 1 comment
#110 - qformer
Issue -
State: closed - Opened by PangziZhang523 28 days ago
- 1 comment
#109 - Finetuning get: ValueError: not enough values to unpack (expected 2, got 0) in get_modality_length_grouped_indices
Issue -
State: open - Opened by sc268 28 days ago
- 2 comments
#108 - Update README.md
Pull Request -
State: closed - Opened by huangleiBuaa about 1 month ago
#107 - Add '--raise_error_at_min_scale False' flag in pretrain and finetune scripts to avoid minimum loss scale crash caused by bad batch.
Issue -
State: closed - Opened by RobotiX101 about 1 month ago
#106 - Tinyllava visualizer
Pull Request -
State: closed - Opened by lingcco about 1 month ago
#105 - 在使用hf上下载TinyLLaVA-Phi-2-SigLIP-3.1B模型时 ,参数设置。
Issue -
State: open - Opened by eva10084 about 1 month ago
#104 - Continue fine-tuning from previously trained model: error on inference
Issue -
State: open - Opened by simoneriggi about 1 month ago
- 8 comments
#103 - Does the Phi template work for Phi3?
Issue -
State: open - Opened by NyKxo1 about 1 month ago
- 1 comment
#102 - About share recipe. Should change "tune_type_vision_tower" to set to "partially-tune"?
Issue -
State: open - Opened by NyKxo1 about 1 month ago
- 1 comment
#101 - About "conv_version" and PretrainTemplate .
Issue -
State: closed - Opened by NyKxo1 about 2 months ago
- 1 comment
#100 - TypeError: __init__() takes 1 positional argument but 2 were given
Issue -
State: closed - Opened by RobotiX101 about 2 months ago
#99 - Exception: Current Loss Scale Already at Minimum While Fine-Tuning with 8 A800-80G GPUs
Issue -
State: closed - Opened by codefanw about 2 months ago
- 2 comments
#98 - add code for qwen2 0.5B
Pull Request -
State: closed - Opened by bhneo 2 months ago
#97 - Can you reproduce the Llava results in this project?
Issue -
State: open - Opened by Yangr116 2 months ago
- 2 comments
#96 - Gemma template for integrating instruct tuned model
Issue -
State: closed - Opened by Jayantverma2 2 months ago
- 1 comment
#95 - replace llm
Issue -
State: open - Opened by LilDevsy0117 2 months ago
- 1 comment
#94 - mof_mlp error
Issue -
State: open - Opened by Daming-W 3 months ago
- 1 comment
#93 - distributed computing
Issue -
State: open - Opened by 1764758458 3 months ago
- 1 comment
#92 - Results are not reproducible on Qwen1.5-1.8B
Issue -
State: open - Opened by Fantasy1120 3 months ago
#91 - dataset seems imcomplete
Issue -
State: closed - Opened by BroJunn 3 months ago
- 1 comment
#90 - Fix `group_by_modality_length` in sync with LLaVA repo
Pull Request -
State: open - Opened by ggcr 3 months ago
#89 - Why `group_by_modality` is set to False on the `custom_finetune.sh` script?
Issue -
State: open - Opened by ggcr 3 months ago
- 3 comments
#88 - Missing key "lm_head.weight" in GemmaForCausalLM when loading lora finetuned TinyLLaVA-Gemma-SigLIP-2.4B
Issue -
State: open - Opened by Yuki-Kokomi 3 months ago
- 5 comments
#87 - Create a tinyllava with qwen2-0.5B-instruct
Issue -
State: open - Opened by bil-ash 3 months ago
#86 - really_cleaned_share-captioner_coco_lcs_sam_1246k_1107.json not found
Issue -
State: open - Opened by Wangman1 3 months ago
- 2 comments
#85 - demo link is down.
Issue -
State: closed - Opened by cooleel 3 months ago
- 1 comment
#84 - ADD SERVE
Pull Request -
State: closed - Opened by jiajunlong 3 months ago
#83 - Questions about LLM templates
Issue -
State: open - Opened by ShawnAn-WHU 3 months ago
- 1 comment
#82 - Reproduced results
Issue -
State: open - Opened by Fantasy1120 3 months ago
- 3 comments
#81 - require_grad
Issue -
State: open - Opened by 1764758458 3 months ago
- 6 comments
#80 - share training recipe
Issue -
State: closed - Opened by Fantasy1120 3 months ago
- 1 comment
#79 - How to inference multiple images from custom finetuned data?
Issue -
State: closed - Opened by hengtseChou 3 months ago
- 2 comments
#78 - --fp16 True question
Issue -
State: open - Opened by Liavan0122 3 months ago
- 7 comments
#77 - template differences?
Issue -
State: open - Opened by tsw123678 3 months ago
- 2 comments
#76 - An idea of using explicit YAML files for configurations
Pull Request -
State: closed - Opened by zjysteven 3 months ago
- 1 comment
#75 - Data about evaluation
Issue -
State: closed - Opened by Fantasy1120 3 months ago
- 2 comments
#74 - Implementing RLHF trainer ?
Issue -
State: open - Opened by R3xpook 3 months ago
- 3 comments
#73 - Does Tiny LLaVa support multi-turn and multi-images conversation?
Issue -
State: open - Opened by Liuziyu77 4 months ago
- 1 comment
#72 - Mistake in eval_textvqa
Issue -
State: open - Opened by hedes1992 4 months ago
- 3 comments
#71 - Question about 'fix everything'
Issue -
State: closed - Opened by Zeqing-Wang 4 months ago
- 3 comments
#70 - Config for TinyLLaVA-OpenELM-270M-SigLIP-0.55B
Issue -
State: open - Opened by eternalding 4 months ago
- 4 comments
#69 - What does "share" mean in the last line of the model performance table.
Issue -
State: open - Opened by lijiannuist 4 months ago
- 1 comment
#68 - Replace vision tower with DINOv2
Issue -
State: closed - Opened by Daming-W 4 months ago
- 1 comment
#67 - consuming time for each VLM pretrain/finetune
Issue -
State: closed - Opened by stark-cs 4 months ago
- 2 comments
#66 - MM-Vet download link broken
Issue -
State: open - Opened by ryan-caesar-ramos 4 months ago
- 2 comments
#65 - Is share_textvqa is the same as textvqa in llava dataset?
Issue -
State: open - Opened by cooleel 4 months ago
- 1 comment
#64 - Error when using the OpenELM as the LLM backbone
Issue -
State: closed - Opened by Zeqing-Wang 4 months ago
- 5 comments
#63 - ModuleNotFoundError: No module named 'tinyllava.model.builder'
Issue -
State: closed - Opened by dkjung 4 months ago
- 3 comments
#62 - Fine-tuning TinyLLaVA-Phi-2-SigLIP-3.1B
Issue -
State: closed - Opened by saadi297 4 months ago
- 8 comments
#61 - Request for Weights of apple/OpenELM-450M-Instruct Model
Issue -
State: closed - Opened by rotem154154 4 months ago
- 3 comments
#60 - Update LICENSE
Pull Request -
State: closed - Opened by shiym2000 4 months ago
#59 - add phi3
Issue -
State: open - Opened by Jayantverma2 4 months ago
#58 - question about custom fine-tuning
Issue -
State: open - Opened by josephtey 4 months ago
#57 - Missing `llavabench.sh`
Issue -
State: open - Opened by cooleel 4 months ago
#56 - Lora finetune results drop dramatically
Issue -
State: closed - Opened by YFCYFC 4 months ago
- 2 comments
#55 - 请问哪行代码在预训练阶段冻结vision tower,好像没看到显式的requires_grad=False
Issue -
State: closed - Opened by Yang-bug-star 4 months ago
#54 - 请问qwen应该将--version 设为mpt吗还是plain?
Issue -
State: open - Opened by Yang-bug-star 4 months ago
- 1 comment
#53 - Loss
Issue -
State: open - Opened by Yu-xm 5 months ago
- 2 comments
#52 - Any plans on training a new model based on Phi-3?
Issue -
State: open - Opened by TheBobbyliu 5 months ago
- 1 comment
#51 - 请问单卡4090能对TinyLLaVA-1.5B进行增量预训练吗?
Issue -
State: open - Opened by Lyzz123 5 months ago
- 4 comments
#50 - Able to merge 1.5B model, but unable to run eval
Issue -
State: open - Opened by tanveer-sayyed 5 months ago
- 6 comments
#49 - Problem about ShareGPT4V SFT supervised finetune dataset
Issue -
State: closed - Opened by YFCYFC 5 months ago
- 4 comments
#48 - can not get the same GQA and MME results using the official TinyLLaVA-1.5B weights
Issue -
State: closed - Opened by YFCYFC 5 months ago
- 2 comments
#47 - Can you share the Loss Curve?
Issue -
State: open - Opened by necrophagists 5 months ago
- 1 comment
#46 - 请问本代码能否在单卡4090上做微调呢?最少显存占用多少
Issue -
State: open - Opened by hby1997 5 months ago
- 1 comment
#45 - Compute budget for TinyLLaVA?
Issue -
State: closed - Opened by ryan-caesar-ramos 6 months ago
- 4 comments
#44 - Some weights of the model checkpoint at bczhou/TinyLLaVA-1.5B
Issue -
State: closed - Opened by dlwlgus53 6 months ago
- 1 comment
#43 - Data download
Issue -
State: open - Opened by wangfengjuan 6 months ago
- 3 comments
#42 - Can you share the finetune.sh and pretrain.sh to train TinyLLaVA-1.5B?
Issue -
State: open - Opened by fyting 6 months ago
- 2 comments
#41 - SigLIP image processing flow
Issue -
State: open - Opened by eternalding 6 months ago
#40 - Fine-tuning on A6000 GPU
Issue -
State: closed - Opened by soham-joshi 6 months ago
- 1 comment
#39 - eval problem
Issue -
State: open - Opened by UestcJay 6 months ago
- 1 comment
#38 - loss always zero
Issue -
State: closed - Opened by ovvo20 6 months ago
- 8 comments
#37 - Is it possible to pretrain tinyllama-3b on 2 V100s ?
Issue -
State: open - Opened by Yang-bug-star 6 months ago
- 1 comment
#36 - llama.cpp support
Issue -
State: closed - Opened by TheBobbyliu 6 months ago
#35 - Saving vision tower in checkpointing
Issue -
State: closed - Opened by zhangliang-04 6 months ago
- 2 comments
#34 - Did you train the whole LLM in the pretraining stage of share recipe?
Issue -
State: open - Opened by Yang-bug-star 6 months ago
- 3 comments
#33 - 请问phi-2是像llava-phi那样预先在文本指令上微调过的吗,还是直接拿基座模型连接视觉编码器进行预训练呢?
Issue -
State: open - Opened by Yang-bug-star 6 months ago
- 1 comment
#32 - Question about commercial usage availability
Issue -
State: closed - Opened by TheBobbyliu 6 months ago
- 1 comment
#31 - Fine-tuning TinyLLaVA-1.5B on custom text-VQA dataset.
Issue -
State: closed - Opened by soham-joshi 6 months ago
- 10 comments
#30 - No module named 'tinyllava.eval.m4c_evaluator'
Issue -
State: closed - Opened by soham-joshi 6 months ago
- 7 comments
#29 - Inference seems can't work
Issue -
State: closed - Opened by JalorOo 6 months ago
- 1 comment
#28 - Gradio demo
Pull Request -
State: closed - Opened by Lealand23231513 6 months ago
#27 - Gradio demo
Pull Request -
State: closed - Opened by Lealand23231513 6 months ago
#26 - Changes to commit:
Pull Request -
State: closed - Opened by Lealand23231513 6 months ago
- 1 comment
#25 - fix: delete the print function in phi.py
Pull Request -
State: closed - Opened by JalorOo 6 months ago
#24 - Use merge_lora_weights.py Error:AttributeError: 'TinyLlavaPhiConfig' object has no attribute 'attention_bias'
Issue -
State: open - Opened by mengjiexu 6 months ago
- 3 comments
#23 - Can Tiny Llama 1.5B lora finetune in one 24GB gpu?
Issue -
State: closed - Opened by JalorOo 6 months ago
- 8 comments
#22 - cls not defined!
Issue -
State: closed - Opened by JalorOo 6 months ago
- 2 comments
#21 - ModuleNotFoundError: No module named 'tinyllava'
Issue -
State: closed - Opened by JalorOo 6 months ago
- 3 comments
#20 - WARNING: tokenization mismatch: 0 vs
Issue -
State: closed - Opened by G-JWLee 6 months ago
- 1 comment
#19 - what the difference between v1 and v1.1?
Issue -
State: open - Opened by Eric-is-good 6 months ago
- 1 comment
#18 - Is it convenient to provide some JSON files
Issue -
State: closed - Opened by wangfengjuan 6 months ago
- 8 comments
#17 - Why is the forward method of the Vision Encoder decorated with @torch.no_grad?
Issue -
State: closed - Opened by Kend-AIL 7 months ago
- 2 comments
#16 - Need training script
Issue -
State: open - Opened by dddraxxx 7 months ago
- 4 comments