Update README.md

README.md

@@ -42,24 +42,3 @@ JAX is extremely powerful, but scaling real training/inference workloads can sti
 - **Production inference**: a dedicated serving stack built for throughput and low latency.
 - **Interoperability**: straightforward workflows with Hugging Face models and assets.
 - **Hackability**: implementations you can actually read, debug, and modify.
-
-## Core components
-
-- **Model library (70+ architectures)**: text models, vision-language models, speech models (Whisper), and more.
-- **Training & fine-tuning**: supervised fine-tuning plus multiple alignment paradigms (preference optimization and RL-style training), and distillation workflows.
-- **eSurge (serving engine)**: continuous batching, paged KV cache, and OpenAI-compatible APIs (plus multimodal endpoints where applicable).
-- **eLargeModel (ELM)**: a configuration-driven interface for end-to-end workflows (load → shard → train → evaluate → serve).
-- **Conversion utilities**: convert PyTorch checkpoints to EasyDeL checkpoints, and generate model cards/metadata during export.
-
-## Where it fits
-
-EasyDeL is a good fit if you:
-
-- want to train or serve large models on JAX with real sharding strategies,
-- need a unified training + serving stack rather than a pile of scripts,
-- care about performance but also want to iterate quickly and customize internals,
-- want compatibility with common Hugging Face assets and workflows.
-
-## License
-
-EasyDeL is released under the Apache-2.0 license.