ise-uiuc/magicoder issues and pull requests

#44 - Huge performace drop when using DeepSpeed-ZeRO2 on 4 A40 GPUs.

Issue - State: open - Opened by twelfth-star 4 months ago

#43 - Question about the different replication result

Issue - State: open - Opened by riddlevv 5 months ago

#43 - Question about the different replication result

Issue - State: open - Opened by riddlevv 5 months ago

#42 - Inquiry about Paper Details of Magicoder

Issue - State: open - Opened by Alex-HaochenLi 7 months ago

#42 - Inquiry about Paper Details of Magicoder

Issue - State: open - Opened by Alex-HaochenLi 7 months ago

#41 - Code for the evaluations on APPS.

Issue - State: open - Opened by wyt2000 8 months ago

#41 - Code for the evaluations on APPS.

Issue - State: open - Opened by wyt2000 8 months ago

#40 - Overlap between Magicoder-Evol-Instruct-110K and HumanEval

Issue - State: closed - Opened by Truth-In-Lies 8 months ago - 2 comments

#40 - Overlap between Magicoder-Evol-Instruct-110K and HumanEval

Issue - State: closed - Opened by Truth-In-Lies 8 months ago - 2 comments

#39 - A question of the generated data from the starcoderdata

Issue - State: closed - Opened by ZJUFLS 9 months ago - 2 comments

#39 - A question of the generated data from the starcoderdata

Issue - State: closed - Opened by ZJUFLS 9 months ago - 2 comments

#38 - Quantised Finetuning on 22GB*4 GPUs

Issue - State: closed - Opened by nikita1503 10 months ago

#38 - Quantised Finetuning on 22GB*4 GPUs

Issue - State: closed - Opened by nikita1503 10 months ago

#37 - HuggingFace Playground has failed

Issue - State: closed - Opened by raminsharifi 11 months ago

#37 - HuggingFace Playground has failed

Issue - State: closed - Opened by raminsharifi 11 months ago

#36 - Is it normal to take more than one hour to get the humanevalplus results?

Issue - State: closed - Opened by yucc-leon 11 months ago

#36 - Is it normal to take more than one hour to get the humanevalplus results?

Issue - State: closed - Opened by yucc-leon 11 months ago

#35 - 8台A40机器上复现magicoder-S-DS-6.7B的结果

Issue - State: closed - Opened by tusiqi1 11 months ago

#35 - 8台A40机器上复现magicoder-S-DS-6.7B的结果

Issue - State: closed - Opened by tusiqi1 11 months ago

#34 - The templates used in reproducing the eval results: why adding the instruction again after "### Response: "?

Issue - State: closed - Opened by yucc-leon 11 months ago - 1 comment

#34 - The templates used in reproducing the eval results: why adding the instruction again after "### Response: "?

Issue - State: closed - Opened by yucc-leon 11 months ago - 1 comment

#33 - catastrophic forgetting problem

Issue - State: closed - Opened by shatealaboxiaowang 11 months ago - 1 comment

#33 - catastrophic forgetting problem

Issue - State: closed - Opened by shatealaboxiaowang 11 months ago - 1 comment

#32 - A scaling law of instruction-code-data would be very interesting...

Issue - State: closed - Opened by yucc-leon 11 months ago - 3 comments
Labels: discussion

#32 - A scaling law of instruction-code-data would be very interesting...

Issue - State: closed - Opened by yucc-leon 11 months ago - 3 comments
Labels: discussion

#31 - Achieved close performance of MagicoderS by finetuning only with `evol-codealpaca-v1`.

Issue - State: closed - Opened by wyt2000 12 months ago - 8 comments
Labels: discussion

#31 - Achieved close performance of MagicoderS by finetuning only with `evol-codealpaca-v1`.

Issue - State: closed - Opened by wyt2000 12 months ago - 8 comments
Labels: discussion

#30 - Got same problem that model only return lots of '\n'

Issue - State: closed - Opened by FoxxComz 12 months ago - 5 comments
Labels: bug

#30 - Got same problem that model only return lots of '\n'

Issue - State: closed - Opened by FoxxComz 12 months ago - 5 comments
Labels: bug

#29 - Data collection and generation

Issue - State: closed - Opened by monaj07 about 1 year ago - 1 comment

#29 - Data collection and generation

Issue - State: closed - Opened by monaj07 about 1 year ago - 1 comment

#28 - Are the training loss and validation loss recorded?

Issue - State: closed - Opened by shatealaboxiaowang about 1 year ago - 4 comments

#28 - Are the training loss and validation loss recorded?

Issue - State: closed - Opened by shatealaboxiaowang about 1 year ago - 4 comments

#27 - How do you set the 'stop_words' parameter

Issue - State: closed - Opened by shatealaboxiaowang about 1 year ago
Labels: discussion, question

#27 - How do you set the 'stop_words' parameter

Issue - State: closed - Opened by shatealaboxiaowang about 1 year ago
Labels: discussion, question

#26 - used Dilated attenton instead of Vanilla Attention in Llama model and fine-tuen the model ,

Issue - State: closed - Opened by younesselbrag about 1 year ago
Labels: help wanted, question

#26 - used Dilated attenton instead of Vanilla Attention in Llama model and fine-tuen the model ,

Issue - State: closed - Opened by younesselbrag about 1 year ago
Labels: help wanted, question

#25 - The correctness of solution

Issue - State: closed - Opened by PilgrimMay about 1 year ago - 1 comment
Labels: documentation, discussion

#25 - The correctness of solution

Issue - State: closed - Opened by PilgrimMay about 1 year ago - 1 comment
Labels: documentation, discussion

#24 - So many impressive experiments ! Are there any experiments with neftune ?

Issue - State: closed - Opened by luoyingyan about 1 year ago - 1 comment
Labels: discussion

#24 - So many impressive experiments ! Are there any experiments with neftune ?

Issue - State: closed - Opened by luoyingyan about 1 year ago - 1 comment
Labels: discussion

#23 - Training data format for Magicoder-OSS-Instruct-75K

Issue - State: closed - Opened by VoiceBeer about 1 year ago - 4 comments
Labels: question

#23 - Training data format for Magicoder-OSS-Instruct-75K

Issue - State: closed - Opened by VoiceBeer about 1 year ago - 4 comments
Labels: question

#22 - The model outputs nothing but "\n" HELP! 😭

Issue - State: closed - Opened by mmmrvector about 1 year ago - 4 comments

#22 - The model outputs nothing but "\n" HELP! 😭

Issue - State: closed - Opened by mmmrvector about 1 year ago - 4 comments

#21 - small models

Issue - State: closed - Opened by BecomeAllan about 1 year ago - 1 comment
Labels: discussion

#21 - small models

Issue - State: closed - Opened by BecomeAllan about 1 year ago - 1 comment
Labels: discussion

#20 - How to write prompt for code completion task

Issue - State: closed - Opened by liHai001 about 1 year ago - 1 comment
Labels: question

#20 - How to write prompt for code completion task

Issue - State: closed - Opened by liHai001 about 1 year ago - 1 comment
Labels: question

#19 - any environment requirement for the model, doesn't work in MacAir M1 (16G)

Issue - State: closed - Opened by larrycai about 1 year ago - 1 comment
Labels: bug, help wanted, question

#19 - any environment requirement for the model, doesn't work in MacAir M1 (16G)

Issue - State: closed - Opened by larrycai about 1 year ago - 1 comment
Labels: bug, help wanted, question

#18 - can I use vLLM to load the models?

Issue - State: closed - Opened by yellowishee about 1 year ago - 1 comment
Labels: question

#18 - can I use vLLM to load the models?

Issue - State: closed - Opened by yellowishee about 1 year ago - 1 comment
Labels: question

#17 - Confusion about the code of train

Issue - State: closed - Opened by jaywongs about 1 year ago - 4 comments
Labels: question

#17 - Confusion about the code of train

Issue - State: closed - Opened by jaywongs about 1 year ago - 4 comments
Labels: question

#16 - the evaluation result of magicoder is not aligned with the result on paper

Issue - State: closed - Opened by WeianMao about 1 year ago - 7 comments
Labels: documentation, question

#16 - the evaluation result of magicoder is not aligned with the result on paper

Issue - State: closed - Opened by WeianMao about 1 year ago - 7 comments
Labels: documentation, question

#15 - can you consider adding my explanation on how to use magicoder in text-generation-webui

Issue - State: closed - Opened by Sairampv98 about 1 year ago - 1 comment
Labels: documentation, enhancement

#15 - can you consider adding my explanation on how to use magicoder in text-generation-webui

Issue - State: closed - Opened by Sairampv98 about 1 year ago - 1 comment
Labels: documentation, enhancement

#14 - Magicoder-S-DS-6.7B gguf, gptq models are not working. Please could you provide quantize models. LM studio, text generation web ui is not is not working.

Issue - State: closed - Opened by MrBenzWorld about 1 year ago - 1 comment
Labels: bug

GitHub / ise-uiuc/magicoder issues and pull requests