-
Notifications
You must be signed in to change notification settings - Fork 152
All issues
Issue creation is restricted in this repository
Issues
is:issue state:open
is:issue state:open
Search results
design(adapters): complete server-mediated adapter binding semantics
area/adapter-functionsGranite adapter functions: framework and adaptiers including RAG, Guardian, CoreGranite adapter functions: framework and adaptiers including RAG, Guardian, Corearea/backendsProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMneeds-designProblem is clear; implementation approach needs design discussion before work startsProblem is clear; implementation approach needs design discussion before work startsp2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1633 In generative-computing/mellea;- Status: Open.#1629 In generative-computing/mellea;
feat: reorder adapter catalog download order
area/adapter-functionsGranite adapter functions: framework and adaptiers including RAG, Guardian, CoreGranite adapter functions: framework and adaptiers including RAG, Guardian, Corep2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1627 In generative-computing/mellea;- Status: Open.#1624 In generative-computing/mellea;
feat: add llmon capabilities for instruct
area/backendsProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMarea/componentComponent type system, CBlock, @generative decorator, prompt templates, RichDocumentComponent type system, CBlock, @generative decorator, prompt templates, RichDocumentarea/formatterOutput formatters, structured output, GraniteFormatter, JSON schema formattingOutput formatters, structured output, GraniteFormatter, JSON schema formattingarea/stdlibCore abstractions: Context, MOT, SamplingStrategy, formatters, serializationCore abstractions: Context, MOT, SamplingStrategy, formatters, serializationp0Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Status: Open.#1623 In generative-computing/mellea;Remove deprecated adapter shims (IntrinsicAdapter/EmbeddedIntrinsicAdapter/CustomIntrinsicAdapter) — PR2 of #1144
area/adapter-functionsGranite adapter functions: framework and adaptiers including RAG, Guardian, CoreGranite adapter functions: framework and adaptiers including RAG, Guardian, Corep2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1621 In generative-computing/mellea;m alora upload's default layout is incompatible with add-readme's generated usage code
area/aloraLoRA/aLora adapter trainingLoRA/aLora adapter trainingarea/clim serve, m alora, CLI commandsm serve, m alora, CLI commandsneeds-designProblem is clear; implementation approach needs design discussion before work startsProblem is clear; implementation approach needs design discussion before work startsp2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1620 In generative-computing/mellea;refactor(backends): consolidate three independent chat_template_kwargs deep-merge implementations
area/backendsProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMp2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1615 In generative-computing/mellea;bug(rag): normalise model-returned JSON types in RAG requirement parsers
bugSomething isn't workingSomething isn't workingp2Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Medium/low: minor bugs, niche features, polish, docs, tests, cleanup. Scoped, lower urgency.Status: Open.#1614 In generative-computing/mellea;Adapter function metrics record success for contract-level IOContract mismatches
area/telemetryOTel spans, metrics, tracing, semconvOTel spans, metrics, tracing, semconvStatus: Open.#1611 In generative-computing/mellea;feat: add think tag parsing for granite models to hugging face backend
area/backendsProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMenhancementNew feature or requestNew feature or requestp0Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Status: Open.feat: add the llmon model id when available
area/backendsProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMProvider-specific work: Ollama, HF, LiteLLM, OpenAI, Bedrock, vLLMp0Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Critical: core-correctness/security breakage, common path, no workaround. Rare emergency bar.Status: Open.