Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • feature/docs
  • feature/models
  • feature/scripts
  • main default protected
  • vladb/e2e_overview
5 results
Created with Raphaël 2.2.016Mar1127Feb2362131Jan30291198431Dec2119161310628Nov19148131Oct9543Added changes to READMEmainmainAdded references to benchmarks in READMEAdded references to benchmarks in READMEAdded code for human-eval benchmarkAdd ARC benchmark evaluation codeAdd 35M model setupMove optimizer state and saved percentage of fp16 activations to globalsFix memory requiremenst, add fp16 activationsUpdate README (fix typos, formatting)Add note about splitting a model's layers on multiple GPU'sAdd compute requirements script and docsAdd more information on memory usage for trainingAdd link to scripts/ in main README fileAdd scripts folder, add memory requirements scriptAdd gradient checkpointing papersfeature/docsfeature/docsAdd GPT1 paperfeature/modelsfeature/modelsUpdate mem/compute scripts interactionfeature/scriptsfeature/scriptsAdd overview of E2E trainingvladb/e2e_overv…vladb/e2e_overviewAdd 35M model setupMove optimizer state and saved percentage of fp16 activations to globalsFix memory requiremenst, add fp16 activationsUpdate README (fix typos, formatting)Add context extension documentationAdd note about splitting a model's layers on multiple GPU'sAdd compute requirements script and docsAdd more information on memory usage for trainingAdd link to scripts/ in main README fileAdd scripts folder, add memory requirements scriptMerge branch 'main-patch-papers' into 'main'Merge branch 'fix/datasets_doc' into 'main'Add link to datasets doc in the README fileMerge branch 'vladb/datasets' into 'main'Add common datasets tableAdd tokenization and mixed precision sectionsMerge branch 'main-patch-3fa8' into 'main'Introduce the Profiling sectionMerge branch 'feature/benchzoo' into 'main'Merge branch 'refactor/structure' into 'main'Add benchzoo repoAdd QLoRA paper, move LoRA paper to fine-tuning
Loading