Skip to content

Issues: ServiceNow/Fast-LLM

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

30 Open 11 Closed
30 Open 11 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[doc] Improve the Checkpoint Conversion Guide for Practical Usability documentation Improvements or additions to documentation
#111 opened Jan 9, 2025 by tscholak
[feat] Implement Loss Masking to Exclude Predefined Token Spans from LM Loss enhancement New feature or request
#109 opened Jan 8, 2025 by tscholak 0.3.0
1
[doc] Benchmark Fast-LLM Pretraining Throughput Against Contemporary Frameworks documentation Improvements or additions to documentation
#108 opened Jan 8, 2025 by tscholak
[doc] Automate MkDocs Documentation for Fast-LLM Configuration Options documentation Improvements or additions to documentation
#107 opened Jan 8, 2025 by tscholak 0.3.0
@tscholak
[meta] Support supervised fine-tuning (SFT) enhancement New feature or request
#106 opened Jan 8, 2025 by tscholak 0.3.0
[feat] Public Non-CLI API for training and data preparation enhancement New feature or request
#105 opened Jan 7, 2025 by tscholak
@bigximik
1
[bug] Incorrect exported config bug Something isn't working
#102 opened Jan 2, 2025 by RaymondLi0
[feat] Track Exact Fast-LLM Version in Training Outputs and wandb Logs enhancement New feature or request
#101 opened Dec 31, 2024 by tscholak
2
[meta] Fast-LLM Improvements Tracker 🌟 enhancement New feature or request
#100 opened Dec 25, 2024 by tscholak 0.3.0
[feat] Improve the weight conversion interface enhancement New feature or request
#99 opened Dec 21, 2024 by jlamypoirier
[feat] Dump full configuration in experiment-dir enhancement New feature or request
#91 opened Dec 10, 2024 by RaymondLi0
2
[bug] Inconsistent init_method_std in test_load_distributed_checkpoint_dp2 bug Something isn't working
#88 opened Dec 9, 2024 by jlamypoirier
@jlamypoirier
[bug] Conversion fails when using layers_per_step with some input formats bug Something isn't working
#87 opened Dec 4, 2024 by RaymondLi0
[feat] Option to disable top-k routing weights normalization enhancement New feature or request
#83 opened Dec 3, 2024 by sohamparikh
1
[feat] QK-Norm enhancement New feature or request
#82 opened Dec 3, 2024 by sohamparikh
[bug] Increasing training loss likely due to desynchronization bug Something isn't working
#77 opened Dec 2, 2024 by tscholak
@nitsanluke @charlesGE
[bug] Barrier timeout for large training runs. bug Something isn't working
#76 opened Nov 29, 2024 by RaymondLi0 0.3.0
@jlamypoirier
[docs] Llama 3.1 8B continual pretraining recipe documentation Improvements or additions to documentation
#70 opened Nov 26, 2024 by tscholak
1
@RaymondLi0
[feat] Support Mamba 2 blocks enhancement New feature or request
#68 opened Nov 25, 2024 by tscholak
[feat] Data lineage tracking in metadata.yaml enhancement New feature or request
#67 opened Nov 25, 2024 by tscholak
@chrish42
[feat] Streamline evaluations: Add integrated Evaluator framework enhancement New feature or request
#65 opened Nov 24, 2024 by tscholak
[feat] FP8 training enhancement New feature or request
#63 opened Nov 22, 2024 by tscholak
[feat] Optional Attention Mask to Prevent Cross-Document Attention in Sequences enhancement New feature or request
#62 opened Nov 22, 2024 by tscholak
[feat] OLMoE hf converter enhancement New feature or request
#61 opened Nov 22, 2024 by tscholak
1
@sohamparikh
[bug] Sparse copy runs out of shared memory with many experts bug Something isn't working
#56 opened Nov 20, 2024 by sohamparikh
@jlamypoirier
1
ProTip! Exclude everything labeled bug with -label:bug.