Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0 #301

dependabot · 2025-01-31T02:29:20Z

Bumps node-llama-cpp from 3.0.0-beta.44 to 3.5.0.

Release notes

Sourced from node-llama-cpp's releases.

v3.5.0

3.5.0 (2025-01-31)

Features

shorter model URIs (#421) (73454d9) (documentation: Model URIs)

Bug Fixes

add missing Jinja features for DeepSeek (#425) (6e4bf3d)

Shipped with llama.cpp release b4600

To use the latest llama.cpp release available, run npx -n node-llama-cpp source download --release latest. (learn more)

v3.4.3

3.4.3 (2025-01-30)

Bug Fixes

adapt to llama.cpp breaking changes (#424) (6e4bf3d)

Shipped with llama.cpp release b4599

To use the latest llama.cpp release available, run npx -n node-llama-cpp source download --release latest. (learn more)

v3.4.2

3.4.2 (2025-01-27)

Bug Fixes

metadata string encoding (#420) (314d7e8)

Vulkan parallel decoding (#420) (314d7e8)

try auth token on 401 response (#420) (314d7e8)

Shipped with llama.cpp release b4567

... (truncated)

Commits

63a1066 chore: update modules (#425)
73454d9 feat: shorter model URIs (#421)
6e4bf3d fix: adapt to llama.cpp breaking changes (#424)
314d7e8 fix: metadata string encoding (#420)
86e1bee fix: ranking empty inputs (#415)
d1b4416 fix: reranking probabilities (#412)
5d07289 chore: update modules (#411)
632a7bf feat: token prediction (speculative decoding) (#405)
e2c5c3f fix: embedding context deadlock (#402)
6a54163 docs: choosing an embedding model (#396)
Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

@dependabot rebase will rebase this PR
@dependabot recreate will recreate this PR, overwriting any edits that have been made to it
@dependabot merge will merge this PR after your CI passes on it
@dependabot squash and merge will squash and merge this PR after your CI passes on it
@dependabot cancel merge will cancel a previously requested merge and block automerging
@dependabot reopen will reopen this PR if it is closed
@dependabot close will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually
@dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
@dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
@dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
@dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

Bumps [node-llama-cpp](https://github.com/withcatai/node-llama-cpp) from 3.0.0-beta.44 to 3.5.0. - [Release notes](https://github.com/withcatai/node-llama-cpp/releases) - [Changelog](https://github.com/withcatai/node-llama-cpp/blob/master/.releaserc.ts) - [Commits](withcatai/node-llama-cpp@v3.0.0-beta.44...v3.5.0) --- updated-dependencies: - dependency-name: node-llama-cpp dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]>

0xdevalias · 2025-02-07T06:20:41Z

@jehna The error in the tests seems to be related to this test:

humanify/src/test/local.e2etest.ts

Lines 14 to 47 in 533cc25

    
           test("Unminifies an example file successfully", async () => { 
        
             const fileIsMinified = async (filename: string) => { 
        
               const prompt = await testPrompt(); 
        
               return await prompt( 
        
                 `Your job is to read the following code and rate it's readabillity and variable names. Answer "EXCELLENT", "GOOD" or "UNREADABLE".`, 
        
                 await readFile(filename, "utf-8"), 
        
                 gbnf`${/("EXCELLENT" | "GOOD" | "UNREADABLE") [^.]+/}.` 
        
               ); 
        
             }; 
        
             const expectStartsWith = (expected: string[], actual: string) => { 
        
               assert( 
        
                 expected.some((e) => actual.startsWith(e)), 
        
                 `Expected "${expected}" but got ${actual}` 
        
               ); 
        
             }; 
        
             await expectStartsWith( 
        
               ["UNREADABLE"], 
        
               await fileIsMinified(`fixtures/example.min.js`) 
        
             ); 
        
             await humanify( 
        
               "local", 
        
               "fixtures/example.min.js", 
        
               "--verbose", 
        
               "--outputDir", 
        
               TEST_OUTPUT_DIR 
        
             ); 
        
             await expectStartsWith( 
        
               ["EXCELLENT", "GOOD"], 
        
               await fileIsMinified(`${TEST_OUTPUT_DIR}/deobfuscated.js`) 
        
             ); 
        
           });

..snip..

# [2025-01-31 02:31:22]  Loading model with options {
#   modelPath: '/Users/runner/.humanifyjs/models/Phi-3.1-mini-4k-instruct-Q4_K_M.gguf',
#   gpuLayers: 0
# }
# Subtest: Unminifies an example file successfully
not ok 4 - Unminifies an example file successfully
  ---
  duration_ms: 85918.501084
  location: '/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:1:468'
  failureType: 'testCodeFailure'
  error: |-
    Expected "UNREADABLE" but got GOOD
    
    The given JavaScript function appears to be well-documented with appropriate variable names and a concise, clear code structure
  code: 'ERR_ASSERTION'
  name: 'AssertionError'
  expected: true
  actual: false
  operator: '=='
  stack: |-
    expectStartsWith (/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:24:5)
    TestContext.<anonymous> (/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:30:9)
    async Test.run (node:internal/test_runner/test:797:9)
    async Test.processPendingSubtests (node:internal/test_runner/test:527:7)
  ...
1..4
# tests 4
# suites 0
# pass 3
# fail 1
# cancelled 0
# skipped 0
# todo 0
# duration_ms 128948.906917
Error: Process completed with exit code 1.

Which is processing this file:

https://github.com/jehna/humanify/blob/533cc25a252aa1827e264b3338dba824d36de741/fixtures/example.min.js

By calling testPrompt:

humanify/src/test/local.e2etest.ts

Lines 16 to 21 in 533cc25

    
           const prompt = await testPrompt(); 
        
           return await prompt( 
        
             `Your job is to read the following code and rate it's readabillity and variable names. Answer "EXCELLENT", "GOOD" or "UNREADABLE".`, 
        
             await readFile(filename, "utf-8"), 
        
             gbnf`${/("EXCELLENT" | "GOOD" | "UNREADABLE") [^.]+/}.` 
        
           );

Which wraps llama:

humanify/src/test/test-prompt.ts

Lines 4 to 8 in 533cc25

    
           export const testPrompt = async () => 
        
             await llama({ 
        
               seed: 1, 
        
               model: process.env["MODEL"] ?? DEFAULT_MODEL 
        
             });

Which returns a function that expects params: systemPrompt, userPrompt, responseGrammar:

humanify/src/plugins/local-llm-rename/llama.ts

Lines 19 to 52 in 533cc25

    
           export async function llama(opts: { 
        
             seed?: number; 
        
             model: string; 
        
             disableGpu?: boolean; 
        
           }): Promise<Prompt> { 
        
             const disableGpu = opts.disableGpu ?? IS_CI; 
        
             const llama = await getLlama({ gpu: disableGpu ? false : "auto" }); 
        
             const modelOpts: LlamaModelOptions = { 
        
               modelPath: getModelPath(opts?.model), 
        
               gpuLayers: disableGpu ? 0 : undefined 
        
             }; 
        
             verbose.log("Loading model with options", modelOpts); 
        
             const model = await llama.loadModel(modelOpts); 
        
             const context = await model.createContext({ seed: opts?.seed }); 
        
             return async (systemPrompt, userPrompt, responseGrammar) => { 
        
               const session = new LlamaChatSession({ 
        
                 contextSequence: context.getSequence(), 
        
                 autoDisposeSequence: true, 
        
                 systemPrompt, 
        
                 chatWrapper: getModelWrapper(opts.model) 
        
               }); 
        
               const response = await session.promptWithMeta(userPrompt, { 
        
                 temperature: 0.8, 
        
                 grammar: new LlamaGrammar(llama, { 
        
                   grammar: `${responseGrammar}` 
        
                 }), 
        
                 stopOnAbortSignal: true 
        
               }); 
        
               session.dispose(); 
        
               return responseGrammar.parseResult(response.responseText); 
        
             }; 
        
           }

The Gbnf helper is passed in as the responseGrammar:

humanify/src/test/local.e2etest.ts

Line 20 in 533cc25

gbnf`${/("EXCELLENT" | "GOOD" | "UNREADABLE") [^.]+/}.`

https://github.com/jehna/humanify/blob/533cc25a252aa1827e264b3338dba824d36de741/src/plugins/local-llm-rename/gbnf.ts

Looking at the node-llama-cpp CHANGELOG since 3.0.0-beta44 up till 3.5.0 for any breaking changes/similar related to the grammar:

https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0-beta.44
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0-beta.45
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0-beta.46
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0-beta.47
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0
- https://node-llama-cpp.withcat.ai/blog/v3
- https://node-llama-cpp.withcat.ai/guide/grammar#using-a-json-schema-grammar
  - Using Grammar
    Use this to enforce a model to generate response in a specific format of text, like JSON for example.
- https://node-llama-cpp.withcat.ai/guide/chat-session#response-json-schema
  - JSON Response With a Schema
- https://node-llama-cpp.withcat.ai/guide/chat-session#function-calling
  - Function Calling
- https://node-llama-cpp.withcat.ai/guide/function-calling
  - Using Function Calling
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.1
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.2
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.3
- adapt to llama.cpp breaking change (fix: adapt to llama.cpp breaking change withcatai/node-llama-cpp#344)
  - Doesn't seem relevant
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.1.0
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.1.1
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.2.0
- adapt to llama.cpp breaking changes (feat: chat session response prefix withcatai/node-llama-cpp#375)
  - Doesn't seem relevant
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.3.0
- adapt to breaking llama.cpp changes (feat: Llama 3.2 3B function calling support withcatai/node-llama-cpp#386)
- Llama 3.2 3B function calling support (feat: Llama 3.2 3B function calling support withcatai/node-llama-cpp#386)
  - Llama3_2LightweightChatWrapper
  - Doesn't seem relevant to the breaking changes, but might be interesting for future functionality
- use llama.cpp backend registry for GPUs instead of custom implementations (feat: Llama 3.2 3B function calling support withcatai/node-llama-cpp#386)
  - Doesn't seem relevant
- getLlama: build: "try" option (feat: Llama 3.2 3B function calling support withcatai/node-llama-cpp#386)
  - "try": If a local build is found, use it. Otherwise, try to build from source and use the resulting binary. If building from source fails, use a prebuilt binary if found.
  - Doesn't seem relevant
- init command: --model flag (feat: Llama 3.2 3B function calling support withcatai/node-llama-cpp#386)
  - Model URI to use. If omitted, you will be prompted to select one interactively
  - Doesn't seem relevant
- JSON Schema grammar: array prefixItems, minItems, maxItems support (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
  - This seems to relate to schemas, which is what is theoretically breaking things currently
  - Mentions GbnfJsonSchema
  - Many changes related to src/utils/gbnfJson/*
- JSON Schema grammar: object additionalProperties, minProperties, maxProperties support (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
  - More schema changes that might be useful for future things, but I don't think are related to the breaking test we're seeing
- JSON Schema grammar: string minLength, maxLength, format support (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
- JSON Schema grammar: improve inferred types (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
  - Some schema related changes, but nothing too dramatic looking
- function calling: params description support (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
  - Some schema related changes, but nothing too dramatic looking
  - Seems to be related to passing a description to the model to assist it with function calling things
- function calling: document JSON Schema type properties on Functionary chat function types (feat: JSON Schema grammar enhancements withcatai/node-llama-cpp#388)
  - Some schema related changes, but nothing too dramatic looking
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.3.1
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.3.2
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.4.0
- token prediction (speculative decoding) (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405) (documentation: Token Prediction)
  - A token predictor is a mechanism that predicts the next few tokens faster than the model can generate them, but the predictions can be inaccurate. We then generate the next token and validate the predictions of the tokens that follow it, all in parallel. After the validation, we discard the incorrect predictions and use the correct ones to speed up the generation process.
    
    Using token predictors doesn't affect the quality of the generated output, but it can speed up the generation process.
- controlledEvaluate (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405) (documentation: Low Level API)
  - To manually control for which of the input tokens to generate output, you can use .controlledEvaluate(...)
- evaluateWithMetadata (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405) (documentation: Low Level API)
  - You can use .evaluateWithMetadata(...) to evaluate tokens onto the context sequence state like .evaluate(...), but with metadata emitted for each token.
- reranking (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405) (documentation: Reranking Documents)
  - Reranking Documents
    After you search for the most similar documents using embedding vectors, you can use inference to rerank (sort) the documents based on their relevance to the given query.
    
    Doing this allows you to combine the best of both worlds: the speed of embedding and the quality of inference.
- token confidence (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405) (documentation: Low Level API)
- experimentalChunkDocument (feat: token prediction (speculative decoding) withcatai/node-llama-cpp#405)
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.4.1
- adapt to breaking llama.cpp changes (fix: ranking empty inputs withcatai/node-llama-cpp#415)
  - Doesn't seem relevant
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.4.2
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.4.3
- adapt to llama.cpp breaking changes (fix: adapt to llama.cpp breaking changes withcatai/node-llama-cpp#424)
  - Doesn't seem relevant
https://github.com/withcatai/node-llama-cpp/releases/tag/v3.5.0
- shorter model URIs (feat: shorter model URIs withcatai/node-llama-cpp#421) (documentation: Model URIs)

0xdevalias · 2025-02-07T11:54:23Z

Looking at the pull request history for bumping node-llama-cpp:

https://github.com/jehna/humanify/pulls?q=sort%3Aupdated-desc+is%3Apr+%22Bump+node-llama-cpp%22

It seems there has been test failures as far back as:

Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0-beta.45 #110
- Unfortunately this was so long ago that the logs have expired and are no longer available: https://github.com/jehna/humanify/actions/runs/10952237439/job/30410578743?pr=110

With the last successful upgrade being:

Bump node-llama-cpp from 3.0.0-beta.42 to 3.0.0-beta.44 #25
- Unfortunately this was so long ago that the logs have expired and are no longer available: https://github.com/jehna/humanify/actions/runs/10342020426/job/28624470785?pr=25

Using a script from my dotfiles to check the old PR checks, and whether they still have logs available:

Old PR Checks

https://github.com/0xdevalias/dotfiles/blob/63d2ab1e0c3fe9e6a6ef4bf89541ab948c32ac90/bin/gh-check-pr-test-logs-exist

⇒ gh-check-pr-test-logs-exist --repo jehna/humanify --dependabot 'node-llama-cpp' --workflow 'Test'
🔍 Using auto-generated Dependabot search query: 'Bump node-llama-cpp author:app/dependabot'

🔍 Fetching PRs in jehna/humanify (Search: 'Bump node-llama-cpp author:app/dependabot')...

📋 Matching PRs:
  PR #25: Bump node-llama-cpp from 3.0.0-beta.42 to 3.0.0-beta.44
  PR #110: Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0-beta.45
  PR #113: Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0-beta.46
  PR #119: Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0
  PR #123: Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.1
  PR #126: Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.3
  PR #148: Bump node-llama-cpp from 3.0.0-beta.44 to 3.1.1
  PR #194: Bump node-llama-cpp from 3.0.0-beta.44 to 3.2.0
  PR #235: Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.0
  PR #243: Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.1
  PR #259: Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.2
  PR #270: Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.0
  PR #289: Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.1
  PR #298: Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.2
  PR #301: Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0

🚀 Fetching workflow checks for each PR (workflow: 'Test')...
  📝 PR #25 - Bump node-llama-cpp from 3.0.0-beta.42 to 3.0.0-beta.44
    SUCCESS Run: 10342020426, Job: 28624470785, Completed: 2024-08-11T18:55:39Z, https://github.com/jehna/humanify/actions/runs/10342020426/job/28624470785
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/10342020426/job/28624470785

  📝 PR #110 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0-beta.45
    FAILURE Run: 10952237439, Job: 30410578743, Completed: 2024-09-20T02:44:38Z, https://github.com/jehna/humanify/actions/runs/10952237439/job/30410578743
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/10952237439/job/30410578743

  📝 PR #113 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0-beta.46
    FAILURE Run: 10986565195, Job: 30500102243, Completed: 2024-09-23T02:21:16Z, https://github.com/jehna/humanify/actions/runs/10986565195/job/30500102243
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/10986565195/job/30500102243

  📝 PR #119 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.0
    FAILURE Run: 11006198263, Job: 30560096048, Completed: 2024-09-24T03:07:11Z, https://github.com/jehna/humanify/actions/runs/11006198263/job/30560096048
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/11006198263/job/30560096048

  📝 PR #123 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.1
    FAILURE Run: 11025072536, Job: 30619327495, Completed: 2024-09-25T02:18:48Z, https://github.com/jehna/humanify/actions/runs/11025072536/job/30619327495
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/11025072536/job/30619327495

  📝 PR #126 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.0.3
    FAILURE Run: 11044698764, Job: 30681065017, Completed: 2024-09-26T02:57:16Z, https://github.com/jehna/humanify/actions/runs/11044698764/job/30681065017
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/11044698764/job/30681065017

  📝 PR #148 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.1.1
    FAILURE Run: 11411055050, Job: 31754595756, Completed: 2024-10-18T21:13:18Z, https://github.com/jehna/humanify/actions/runs/11411055050/job/31754595756
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/11411055050/job/31754595756

  📝 PR #194 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.2.0
    FAILURE Run: 11605087286, Job: 32314887872, Completed: 2024-10-31T02:37:10Z, https://github.com/jehna/humanify/actions/runs/11605087286/job/32314887872
    ❌ Logs expired/unavailable - 🔗 https://github.com/jehna/humanify/actions/runs/11605087286/job/32314887872

  📝 PR #235 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.0
    FAILURE Run: 12131715473, Job: 33824331143, Completed: 2024-12-03T02:29:45Z, https://github.com/jehna/humanify/actions/runs/12131715473/job/33824331143
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/12131715473/job/33824331143
    📋 View failed logs: gh run view 12131715473 --repo jehna/humanify --log-failed

  📝 PR #243 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.1
    FAILURE Run: 12228072358, Job: 34105732535, Completed: 2024-12-09T03:00:43Z, https://github.com/jehna/humanify/actions/runs/12228072358/job/34105732535
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/12228072358/job/34105732535
    📋 View failed logs: gh run view 12228072358 --repo jehna/humanify --log-failed

  📝 PR #259 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.2
    FAILURE Run: 12554593031, Job: 35003489033, Completed: 2024-12-31T02:37:09Z, https://github.com/jehna/humanify/actions/runs/12554593031/job/35003489033
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/12554593031/job/35003489033
    📋 View failed logs: gh run view 12554593031 --repo jehna/humanify --log-failed

  📝 PR #270 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.0
    FAILURE Run: 12663197159, Job: 35289351219, Completed: 2025-01-08T02:58:25Z, https://github.com/jehna/humanify/actions/runs/12663197159/job/35289351219
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/12663197159/job/35289351219
    📋 View failed logs: gh run view 12663197159 --repo jehna/humanify --log-failed

  📝 PR #289 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.1
    FAILURE Run: 12941994481, Job: 36098955172, Completed: 2025-01-24T02:26:22Z, https://github.com/jehna/humanify/actions/runs/12941994481/job/36098955172
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/12941994481/job/36098955172
    📋 View failed logs: gh run view 12941994481 --repo jehna/humanify --log-failed

  📝 PR #298 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.2
    FAILURE Run: 13002502354, Job: 36263639435, Completed: 2025-01-28T02:39:24Z, https://github.com/jehna/humanify/actions/runs/13002502354/job/36263639435
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/13002502354/job/36263639435
    📋 View failed logs: gh run view 13002502354 --repo jehna/humanify --log-failed

  📝 PR #301 - Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0
    FAILURE Run: 13065242932, Job: 36456364632, Completed: 2025-01-31T02:32:51Z, https://github.com/jehna/humanify/actions/runs/13065242932/job/36456364632
    ✅ Logs exist: https://github.com/jehna/humanify/actions/runs/13065242932/job/36456364632
    📋 View failed logs: gh run view 13065242932 --repo jehna/humanify --log-failed

The oldest PR that still does have test error logs available seems to be:

Bump node-llama-cpp from 3.0.0-beta.44 to 3.3.0 #235

https://github.com/jehna/humanify/actions/runs/12131715473/job/33824331143

..snip..

# [2024-12-03 02:28:15]  Loading model with options {
#   modelPath: '/Users/runner/.humanifyjs/models/Phi-3.1-mini-4k-instruct-Q4_K_M.gguf',
#   gpuLayers: 0
# }
# Subtest: Unminifies an example file successfully
not ok 4 - Unminifies an example file successfully
  ---
  duration_ms: 87609.07725
  location: '/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:1:468'
  failureType: 'testCodeFailure'
  error: |-
    Expected "UNREADABLE" but got GOOD
    
    The provided function `a` appears to be a sublist generator for slicing a string or an array in JavaScript
  code: 'ERR_ASSERTION'
  name: 'AssertionError'
  expected: true
  actual: false
  operator: '=='
  stack: |-
    expectStartsWith (/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:24:5)
    TestContext.<anonymous> (/Users/runner/work/humanify/humanify/src/test/local.e2etest.ts:30:9)
    async Test.run (node:internal/test_runner/test:797:9)
    async Test.processPendingSubtests (node:internal/test_runner/test:527:7)
  ...
1..4
# tests 4
# suites 0
# pass 3
# fail 1
# cancelled 0
# skipped 0
# todo 0
# duration_ms 125798.5[195](https://github.com/jehna/humanify/actions/runs/12131715473/job/33824331143#step:9:196)41
Error: Process completed with exit code 1.

Looking at the CHANGELOG entry for the version that started breaking:

https://github.com/withcatai/node-llama-cpp/releases/tag/v3.0.0-beta.45

The following links may be relevant for understanding if/what changes need to be made to the Gbnf grammar (or other ways to achieve similar):

dependabot bot added the dependencies Pull requests that update a dependency file label Jan 31, 2025

dependabot bot mentioned this pull request Jan 31, 2025

Bump node-llama-cpp from 3.0.0-beta.44 to 3.4.2 #298

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0 #301

Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0 #301

dependabot bot commented on behalf of github Jan 31, 2025

0xdevalias commented Feb 7, 2025

0xdevalias commented Feb 7, 2025 •

edited

Loading

Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0 #301

Are you sure you want to change the base?

Bump node-llama-cpp from 3.0.0-beta.44 to 3.5.0 #301

Conversation

dependabot bot commented on behalf of github Jan 31, 2025

v3.5.0

3.5.0 (2025-01-31)

Features

Bug Fixes

v3.4.3

3.4.3 (2025-01-30)

Bug Fixes

v3.4.2

3.4.2 (2025-01-27)

Bug Fixes

0xdevalias commented Feb 7, 2025

0xdevalias commented Feb 7, 2025 • edited Loading

0xdevalias commented Feb 7, 2025 •

edited

Loading