Search results

server : separate the notion of position and KV tokens, remove prompt truncation

breaking changeChanges that break ABIs, APIs, file formats, or other forms of backwards compatibility.Changes that break ABIs, APIs, file formats, or other forms of backwards compatibility.

examples

pythonpython script changespython script changes

server

Status: Open (in progress).

ggml-org/llama.cppnumber 13576

#13576 In ggml-org/llama.cpp;

· ngxson opened on May 15, 2025

llama : rename batch.logits to batch.output

androidIssues specific to AndroidIssues specific to Android

breaking changeChanges that break ABIs, APIs, file formats, or other forms of backwards compatibility.Changes that break ABIs, APIs, file formats, or other forms of backwards compatibility.

examples

server

Status: Open (in progress).

ggml-org/llama.cppnumber 10004

#10004 In ggml-org/llama.cpp;

· danbev opened on Oct 22, 2024

imatrix : use GGUF to store importance matrices

breaking changeChanges that break ABIs, APIs, file formats, or other forms of backwards compatibility.Changes that break ABIs, APIs, file formats, or other forms of backwards compatibility.

enhancementNew feature or requestNew feature or request

examples

pythonpython script changespython script changes

refactoringRefactoringRefactoring

Review Complexity : MediumGenerally require more time to grok but manageable by beginner to medium expertise levelGenerally require more time to grok but manageable by beginner to medium expertise level

Status: Open (in progress).

ggml-org/llama.cppnumber 9400

#9400 In ggml-org/llama.cpp;

· compilade opened on Sep 10, 2024

<code>json</code>: $ref + object overhaul (https &amp; recursive $refs, mix properties &amp; allOf)

breaking changeChanges that break ABIs, APIs, file formats, or other forms of backwards compatibility.Changes that break ABIs, APIs, file formats, or other forms of backwards compatibility.

examples

pythonpython script changespython script changes

Review Complexity : MediumGenerally require more time to grok but manageable by beginner to medium expertise levelGenerally require more time to grok but manageable by beginner to medium expertise level

server

testingEverything test relatedEverything test related

Status: Open (in progress).

ggml-org/llama.cppnumber 8199

#8199 In ggml-org/llama.cpp;

· ochafik opened on Jun 28, 2024

GGUF endianness cannot be determined from GGUF itself

breaking changeChanges that break ABIs, APIs, file formats, or other forms of backwards compatibility.Changes that break ABIs, APIs, file formats, or other forms of backwards compatibility.

enhancementNew feature or requestNew feature or request

good first issueGood for newcomersGood for newcomers

Status: Open.

#3957 In ggml-org/llama.cpp;

· philpax opened on Nov 5, 2023