Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / microsoft/torchscale issues and pull requests
#113 - torchscale 0.3.0 does not include LongNet
Issue -
State: open - Opened by tsy935 2 months ago
#112 - train.py: error: unrecognized arguments: --flash-attention --segment-length [2048,4096] --dilated-ratio [1,2]
Issue -
State: open - Opened by github2657529567 3 months ago
#111 - No module named 'sentencepiece'
Issue -
State: open - Opened by github2657529567 3 months ago
#110 - torchscale 0.3.0 requires fairscale==0.4.0, but you have fairscale 0.4.13 which is incompatible.
Issue -
State: open - Opened by pandayummy 5 months ago
#109 - Minecraft
Issue -
State: closed - Opened by Pelaez99 7 months ago
#108 - Question about LongNet attention map overlap
Issue -
State: open - Opened by RmZeta2718 9 months ago
#107 - Different batch sizes lead to different evalution results for LongVIT
Issue -
State: open - Opened by HHHedo 9 months ago
#106 - How to test the model
Issue -
State: open - Opened by ReloJeffrey 9 months ago
#105 - pip error
Issue -
State: open - Opened by wanghaoran-ucas 10 months ago
#104 - Where is the offset implemented in Multi-head dilated attention ?
Issue -
State: open - Opened by AshStuff 10 months ago
#103 - can't use longvit
Issue -
State: open - Opened by abebe9849 10 months ago
#102 - Question about learnable segment lengths and dilation rates
Issue -
State: open - Opened by benrousePUC 10 months ago
#101 - How to use retention in RetNet for cross-attention?
Issue -
State: open - Opened by yxchng 11 months ago
#100 - renames longnet file; longnet example in readme works now
Pull Request -
State: open - Opened by JacksonSearle 11 months ago
- 1 comment
#99 - Checkpoint for RetNet
Issue -
State: open - Opened by macsz 11 months ago
#98 - What WSI level was used for pretraining LongVit?
Issue -
State: closed - Opened by jpfeil 12 months ago
- 1 comment
#97 - about attention mask
Issue -
State: closed - Opened by hichoe95 about 1 year ago
#96 - Bump pillow from 10.0.0 to 10.2.0 in /examples/longvit
Pull Request -
State: open - Opened by dependabot[bot] about 1 year ago
Labels: dependencies
#95 - about the longnet's ppl
Issue -
State: open - Opened by robotzheng about 1 year ago
- 2 comments
#94 - Update requirements.txt
Pull Request -
State: closed - Opened by I8dNLo about 1 year ago
#93 - Bump transformers from 4.8.1 to 4.36.0 in /examples/longvit
Pull Request -
State: open - Opened by dependabot[bot] about 1 year ago
Labels: dependencies
#92 - Fix No module named 'torch._six'
Pull Request -
State: open - Opened by ahmedhshahin about 1 year ago
#91 - Bump pyarrow from 9.0.0 to 14.0.1 in /examples/longvit
Pull Request -
State: open - Opened by dependabot[bot] about 1 year ago
Labels: dependencies
#90 - Bump scipy from 1.6.3 to 1.10.0 in /examples/longvit
Pull Request -
State: open - Opened by dependabot[bot] about 1 year ago
Labels: dependencies
#89 - Bump pillow from 10.0.0 to 10.0.1 in /examples/longvit
Pull Request -
State: closed - Opened by dependabot[bot] about 1 year ago
- 1 comment
Labels: dependencies
#88 - Bump transformers from 4.8.1 to 4.30.0 in /examples/longvit
Pull Request -
State: closed - Opened by dependabot[bot] about 1 year ago
- 1 comment
Labels: dependencies
#87 - Release LongNet and LongViT
Pull Request -
State: closed - Opened by shumingma about 1 year ago
#86 - Wrong Rnm Normalization.
Issue -
State: open - Opened by pdradx about 1 year ago
- 1 comment
#85 - Introducing padding_mask to RetNet
Issue -
State: open - Opened by xtwigs about 1 year ago
- 2 comments
#84 - Question regarding the configuration of decoder_retention_heads
Issue -
State: open - Opened by Kratos-Wen about 1 year ago
- 2 comments
#83 - Training RetNet on A100 GPUs
Issue -
State: open - Opened by Antoine-Bergerault about 1 year ago
- 1 comment
#82 - [Minor issue] Discrepancy inside arxiv paper
Issue -
State: open - Opened by radarFudan about 1 year ago
#81 - Question about the normalization in attention
Issue -
State: closed - Opened by Cranial-XIX over 1 year ago
- 2 comments
#80 - Question about RetNetRelPos
Issue -
State: closed - Opened by hyunwoongko over 1 year ago
- 2 comments
#79 - about gamma/decay in RetNet
Issue -
State: closed - Opened by rouniuyizu over 1 year ago
- 2 comments
#78 - typo in normalization denominator in parallel retention?
Issue -
State: closed - Opened by XintianHan over 1 year ago
- 1 comment
#77 - Chunk recurrent representation incorrect results
Issue -
State: closed - Opened by N0r9st over 1 year ago
- 7 comments
#76 - Query about Retentive Network's Recurrent Representation
Issue -
State: closed - Opened by gopi231091 over 1 year ago
- 1 comment
#75 - About training memory
Issue -
State: closed - Opened by HoraceXIaoyiBao over 1 year ago
- 2 comments
#74 - BEiT3 Vision-Language Expert question
Issue -
State: closed - Opened by andreapdr over 1 year ago
- 4 comments
#73 - AttributeError: 'EncoderDecoderConfig' object has no attribute 'normalize_output'
Issue -
State: closed - Opened by 3CodeLove over 1 year ago
- 3 comments
#72 - RuntimeError: The size of tensor a (5) must match the size of tensor b (2) at non-singleton dimension 0
Issue -
State: closed - Opened by codinglover0111 over 1 year ago
- 3 comments
#71 - Compatibility with torchsummary
Issue -
State: closed - Opened by lzqlzzq over 1 year ago
- 1 comment
#70 - fix fairseq example
Pull Request -
State: closed - Opened by sunyt32 over 1 year ago
#69 - Update new RetNet settings
Pull Request -
State: closed - Opened by sunyt32 over 1 year ago
#68 - initialization of qkv
Issue -
State: closed - Opened by XintianHan over 1 year ago
- 3 comments
#67 - pip package does not contain RetNet
Issue -
State: closed - Opened by fabienGenhealth over 1 year ago
- 2 comments
#66 - Question on decay factor for attention with xPos
Issue -
State: closed - Opened by mvbakulin over 1 year ago
- 1 comment
#65 - There're a confusion in torchscale
Issue -
State: closed - Opened by lovekang3344 over 1 year ago
- 3 comments
#64 - retnet traning config
Issue -
State: open - Opened by hanlinxuy over 1 year ago
- 7 comments
#63 - Could you please explain the reason behind defining TEMPERATURE_FOR_L_UAX in the code without actually using it?
Issue -
State: closed - Opened by Ruiyuan-Zhang over 1 year ago
- 1 comment
#62 - Question about the recurrent forward of MultiScaleRetention
Issue -
State: closed - Opened by LEECHOONGHO over 1 year ago
- 2 comments
#61 - Can Torchscale be applied in point cloud tasks?
Issue -
State: closed - Opened by huihui0613 over 1 year ago
- 2 comments
#60 - `get_moe_group` 's return is None, when building `class MOELayer(Base)` , using one gpu
Issue -
State: closed - Opened by Ruiyuan-Zhang over 1 year ago
- 4 comments
#59 - embed_tokens
Issue -
State: open - Opened by CodeMiningCZW over 1 year ago
- 4 comments
#58 - Question about is_first_step and Retnet
Issue -
State: closed - Opened by tdomhan over 1 year ago
- 2 comments
#57 - Retnet parameter dimension
Issue -
State: closed - Opened by allanj over 1 year ago
- 2 comments
#56 - "sentencepiece.bpe.model" and "dict.txt" in page below seem not available
Issue -
State: closed - Opened by HuXinjing over 1 year ago
- 2 comments
#55 - Retnet training is slow
Issue -
State: closed - Opened by Zth9730 over 1 year ago
- 2 comments
#54 - RetNet : Check consistency of each forward mode
Issue -
State: closed - Opened by mmorinag127 over 1 year ago
- 9 comments
#53 - Is there some example of the paper? e.g., compare of the inference latency
Issue -
State: closed - Opened by LiZeng001 over 1 year ago
- 1 comment
#52 - Training & Inference examples for RetNet
Issue -
State: closed - Opened by jhl-Det over 1 year ago
- 1 comment
#51 - fix chunkwise inconsistency bug
Pull Request -
State: closed - Opened by sunyt32 over 1 year ago
#50 - Adding sqrt in the recurrent_forward of retnet to make it consistent with parallel_forward
Pull Request -
State: closed - Opened by wangmengzhi over 1 year ago
#49 - RetNet: relative position
Issue -
State: closed - Opened by fkodom over 1 year ago
- 5 comments
#48 - Multi-Scale Retention: Why include position embeddings explicitly?
Issue -
State: closed - Opened by fkodom over 1 year ago
- 3 comments
#47 - scale.sqrt() in the recurrent_forward function of the multiscale_retention module
Issue -
State: closed - Opened by wangmengzhi over 1 year ago
- 6 comments
#46 - Update epsilon in retention
Pull Request -
State: closed - Opened by sunyt32 over 1 year ago
#45 - LEX inference support and checkpoint
Issue -
State: closed - Opened by RulinShao over 1 year ago
- 5 comments
#44 - recurrent_forward in MultiScaleRetention
Issue -
State: closed - Opened by Anker-ZX-AI over 1 year ago
- 1 comment
#43 - AttributeError: 'EncoderConfig' object has no attribute 'decoder_layers'
Issue -
State: closed - Opened by dedekinds over 1 year ago
- 2 comments
#42 - the meaning of "incremental_state" in RetNet
Issue -
State: closed - Opened by jhl-Det over 1 year ago
- 3 comments
#41 - can not download dict.txt
Issue -
State: closed - Opened by robotzheng over 1 year ago
- 2 comments
#40 - Inconsist recurrent and parallel results for RetNet
Issue -
State: closed - Opened by YirunKCL over 1 year ago
- 4 comments
#39 - Config fix
Pull Request -
State: open - Opened by agoryuno over 1 year ago
#38 - Remove inheritance from `object`
Pull Request -
State: open - Opened by agoryuno over 1 year ago
- 2 comments
#37 - Longnet Code Release
Issue -
State: closed - Opened by arnavdantuluri over 1 year ago
- 13 comments
#36 - testing very large attention windows
Issue -
State: open - Opened by fredzannarbor over 1 year ago
#35 - About the param `scale_base`
Issue -
State: closed - Opened by horizon94 over 1 year ago
- 1 comment
#34 - some result plots are not show
Issue -
State: closed - Opened by klae01 over 1 year ago
- 1 comment
#33 - support lm prefix computation in one go
Pull Request -
State: closed - Opened by XingxingZhang over 1 year ago
#32 - EncoderDecoder Configuration Issue
Issue -
State: closed - Opened by klae01 over 1 year ago
- 1 comment
#31 - add basic test
Pull Request -
State: closed - Opened by klae01 over 1 year ago
- 1 comment
#30 - make pgs global
Pull Request -
State: closed - Opened by njb-ms almost 2 years ago
- 2 comments
#29 - question about the number of output_projection
Issue -
State: closed - Opened by violet-sto almost 2 years ago
- 1 comment
#28 - xPos cross-attention change
Issue -
State: closed - Opened by janEbert almost 2 years ago
- 2 comments
#27 - Bump timm version to latest
Pull Request -
State: closed - Opened by JonathanRayner almost 2 years ago
#25 - Fairseq version compatible with torchscale
Issue -
State: closed - Opened by sjelassi almost 2 years ago
- 1 comment
#24 - Swapped naive dot product attention for flash attention
Pull Request -
State: open - Opened by usryokousha almost 2 years ago
- 4 comments
#23 - About running speed
Issue -
State: open - Opened by NieShenRuc almost 2 years ago
#22 - Could not install fairseq
Issue -
State: closed - Opened by BaohaoLiao almost 2 years ago
- 1 comment
#21 - v0.2.0
Pull Request -
State: closed - Opened by shumingma almost 2 years ago
#20 - fx BERT + moe
Pull Request -
State: closed - Opened by buaahsh almost 2 years ago
#19 - Update README.md
Pull Request -
State: closed - Opened by buaahsh almost 2 years ago
#18 - Installer bug - wrong `apex` package installed
Issue -
State: closed - Opened by jph00 almost 2 years ago
- 2 comments
#17 - SMOE or XMOE Network how to "evaluate" and "save and resume"
Issue -
State: closed - Opened by randomtutu almost 2 years ago
- 2 comments
#16 - Questions about the implementation of deepnorm
Issue -
State: closed - Opened by jiaohuix almost 2 years ago
- 2 comments
#15 - [Question] what are the usages of multiway_network.py?
Issue -
State: closed - Opened by berniewang8177 almost 2 years ago
- 2 comments
#14 - Does Torchscale support vision transformers in vision tasks?
Issue -
State: closed - Opened by nightsnack about 2 years ago
- 5 comments
#13 - Q) Tensor parallel for magneto
Issue -
State: closed - Opened by taehwakkwon about 2 years ago
- 9 comments