Skip to content
Forgather
Mistral
Initializing search
jdinalt/forgather
Home
Getting Started
Core Concepts
Configuration
Training
Checkpointing
Models & Inference
Datasets
Fused Loss
Guides
API Reference
Tutorials
Development
Release Notes
Forgather
jdinalt/forgather
Home
Getting Started
Getting Started
Overview
Installation
Docker images
Core Concepts
Configuration
Configuration
Overview
Syntax Reference
Model Initialization
Debugging
Low-level API
Project Templates
Training
Training
Trainer Options
Pipeline Parallel
Trainer Control
Performance Metrics
DiLoCo
FP8 Training
TorchTitan
Adafactor Triton
Checkpointing
Checkpointing
Overview
User Guide
Distributed Abstraction
Migration Guide
Sharded Checkpoint API
Divergence Detection
Models & Inference
Models & Inference
Model Architecture
Model Conversion
Finalize Model
Add-Tokens Config
EOS and generate() Stopping
vLLM Integration
Datasets
Datasets
Sequence Packing
Quick Reference
Document Boundaries
Fast HF Loader
Fast HF Loader Checkpoints
Dataset Server
Dataset Projects
Dataset CLI
Fused Loss
Fused Loss
Cross-Entropy Comparison
Trainer API
Apple CCE Analysis
Pipeline Integration
Guides
Guides
Creating a Model Project
Model CLI Reference
Creating a Dataset Project
Working with Tokenizer Projects
Debugging
Interactive CLI
Evaluating Models
Log Analysis
TensorBoard
MkDocs
API Reference
API Reference
Overview
Project System
Trainers
Callbacks
Checkpoints
Optimizers
Datasets
Analysis
Tutorials
Tutorials
Tiny Llama
HP Lovecraft Project
Samantha Finetune
Forgather Projects Overview
Project Index
Dynamic LM
Development
Development
Testing
Integration Testing
Git Hooks
Known Bugs
Release Notes
Release Notes
Overview
1.2.1
1.2.0
Mistral
¶
Mistral with sliding window attention support
Back to top