foundation-model-stack fms-acceleration issues

foundation-model-stack / fms-acceleration

🚀 Collection of libraries used with fms-hf-tuning to accelerate fine-tuning and training of large models.

Apache License 2.0

6 stars 12 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Enable GPTQModel to handle GraniteMoeParallelExperts

#112 fabianlim opened 6 days ago
0
fix: requirements file path in error

#111 willmj closed 1 week ago
0
Slow down observed for BigCode Santa Coder

#110 fabianlim opened 1 week ago
0
ScatterMoE Gradient Norm Needs to be Properly Computed When Used With FSDP

#109 fabianlim opened 2 weeks ago
0
Improve Documentation

#108 fabianlim opened 2 weeks ago
0
FIx Deprecation Warnings in AutoCast API

#107 fabianlim opened 2 weeks ago
0
Disable MLP Fused Ops if Not SwiGLU, Depracted Fast Quantized Peft Plugin, Update Benchmarks

#106 fabianlim closed 1 week ago
0
Extract ScatterMoE TriTon Kernels from Kernel Hyperdrive Fork

#105 fabianlim opened 2 weeks ago
0
Incorporate Liger

#104 fabianlim opened 2 weeks ago
0
ScatterMoE to support LoRA Adapters

#103 fabianlim opened 2 weeks ago
0
Fix Dropout in Fused LoRA Operations

#102 fabianlim closed 2 weeks ago
0
ScatterMoE to support Quantized PEFT

#101 fabianlim opened 2 weeks ago
0
Numba JIT TypingErrors Thrown on Multipack Functions

#100 fabianlim opened 2 weeks ago
0
Add ExpertParallel Mixture-of-Experts Plugin

#99 fabianlim closed 1 week ago
0
FOAK Cross Entropy Loss Will Not Work with New Loss Functions After Transformers 4.46

#98 fabianlim opened 4 weeks ago
1
Dropout for LoRA Fused Ops Incorrectly Implemented. Also should be more careful about activation

#97 fabianlim closed 2 weeks ago
0
Fix Issue with Resizing Parameters on the Meta Device in Low CPU Mem Mode

#96 fabianlim closed 4 weeks ago
1
model: Add granite GPTQ model

#95 willmj closed 3 weeks ago
4
Apply Retie Weights Fix Regardless of Transformers and TRL version for AutoGPTQ

#94 fabianlim closed 1 month ago
0
feat: add liger kernel with fused cross entropy loss

#93 anhuong opened 1 month ago
7
Quickfix: Accelerate YAML and LoRA Fused Ops

#92 fabianlim closed 1 month ago
0
Register Kernels as AutoGrad Ops

#91 fabianlim opened 1 month ago
1
Fix Low CPU Memory Mode Issues for Quantized Peft

#90 fabianlim closed 1 month ago
0
Fixes to Accelerated Peft

#89 fabianlim closed 1 month ago
2
Expand support for Collator caller functions

#88 kmehant closed 1 month ago
2
Support for Iterable Datasets

#87 kmehant closed 3 weeks ago
2
Update Benchmarks and Documentation for GraniteCausalLM

#86 fabianlim closed 1 month ago
3
Update Benches: Orca

#85 fabianlim closed 1 month ago
0
Slowdown and Higher Memory Consumption for GPTQ-LoRA with Bfloat16

#84 achew010 opened 2 months ago
1
Distributed Training Problems for QLoRA models with Transformers pre-release 4.45

#83 achew010 closed 1 month ago
2
Fix Bug on Peft Config Check in AutoGPTQ Plugin

#82 achew010 closed 2 months ago
0
fixed bug in peft installation for gptqmodel

#81 achew010 closed 2 months ago
2
Create fused_CEloss.py

#80 wynterl closed 1 month ago
1
Allow Kernels for Full FT and Non-Quantized PEFT

#79 fabianlim closed 2 months ago
2
Allow PaddingFree to work with DataCollatorForCompletionOnlyLM

#78 fabianlim closed 2 months ago
1
Ensure Model is Correctly Loaded for Augmentation Purposes

#77 fabianlim opened 2 months ago
0
Introduce Liger Fused Cross Entropy Kernel to FOAK Plugin

#76 achew010 opened 2 months ago
3
When HF Memory Metrics Disabled, the Benchmark CSV is Corrupted.

#75 fabianlim closed 1 month ago
1
Fix formatter

#74 achew010 closed 3 months ago
0
fix: bounds and logger

#73 kmehant closed 3 months ago
0
fix: bounds and logger

#72 kmehant closed 3 months ago
0
Simplifications to MultiPack Plugin and Unit Tests

#71 achew010 closed 3 months ago
0
Inconsistency in Padding-Free Benchmarks with Different Transformers Versions

#70 achew010 opened 3 months ago
3
Mixture of Experts Training with Acceleration Library Plugin

#69 fabianlim closed 1 week ago
1
Benchmarking: Add Response Field to Use Chat Templates Without Response Template

#68 fabianlim closed 3 months ago
0
Add Acceleration Patcher and MultiPack Plugin

#67 fabianlim closed 3 months ago
2
Add Benchmarking Compatibility to PaddingFree Plugin

#66 achew010 closed 3 months ago
2
Shift Computation of PaddingFree Variable `CuSeqLen` from Flash Attention Forward to DataCollatorWithFlattening

#65 achew010 closed 3 months ago
1
Rename Plugin to `AttentionAndDistributedPacking `

#64 achew010 closed 3 months ago
1
Rectify Missing Dataloader Preparation Call in PaddingFree Plugin Method

#63 achew010 closed 3 months ago
1