snapshot-2024-02-11

What's Changed

• Bump llama-cpp-python to 0.2.38 by @oobabooga in #5420
• Quadratic sampling by @kalomaze in #5403
• Remove non-HF ExLlamaV2 loader by @oobabooga in #5431
• Fix the n_batch slider by @BadisG in #5436
• Split by rows instead of layers for llama.cpp multi-gpu by @Ph0rk0z in #5435
• Improve ChatML template by @BadisG in #5411
• Truncate long chat completions inputs by @oobabooga in #5439
• Add custom sampler order support by @oobabooga in #5443
• Merge dev branch by @oobabooga in #5452
• Merge dev branch by @oobabooga in #5453

Full Changelog: snapshot-2024-02-04...snapshot-2024-02-11