14 versions

• Added adapter_source and api_token to completions API (#446)○416c3adc2024-04-29 22:59
• Make max_new_tokens optional, default to max_total_tokens - input_length (#353)○68ff0bf52024-03-22 17:01
• enh: Expose ignore_eos_token option in generate requests (#340)○367579cc2024-03-19 21:09
• Added Outlines logits processor for JSON schema validation (#224)○3309618c2024-02-12 23:38
• Merge multiple LoRA adapters per request (linear, TIES, DARE) (#212)○33b4c9732024-02-01 05:39
• Fixed OpenAI stream response data (#193)●2○189062532024-01-19 06:11
• fix: OpenAI response format (#184)▲1●2○376fcc5d2024-01-16 16:48
• docs: OpenAI compatible API (#174)○264739ad2024-01-10 19:29
• OpenAI v1 Chat Completions API (#171)○3a90d4432024-01-10 18:35
• Added `prompt_tokens` to the response (#165)○4d88ffed2024-01-09 05:54
• Add predibase as a source for adapters (#125)○3143f3032023-12-14 23:55
• Update openapi.jsonno API changes41c657e2023-12-07 19:33
• Mkdocs (#112)▲9●18○41284aaa22023-12-07 18:43
• 0.5.0latest2026-07-17 12:12