github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #9202] [CLOSED] model: add a test for model forward pass during implementation #38760

New Issue
Closed
opened 2026-04-22 23:25:35 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-22 23:25:35 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/9202
Author: @BruceMacD
Created: 2/18/2025
Status: ❌ Closed

Base: main ← Head: brucemacd/forward-test

📝 Commits (1)

  • 31d04eb model: add a test for model forward pass during implementation

📊 Changes

4 files changed (+158 additions, -0 deletions)

View changed files

📝 .gitignore (+3 -0)
➕ model/model_external_test.go (+138 -0)
➕ model/testdata/models/README.md (+10 -0)
➕ model/testdata/models/qwen2_5.json (+7 -0)

📄 Description

This PR introduces a new testing framework for validating model forward pass behavior using JSON-specified test cases. It helps developers verify their model implementations by running simple prompts through the model and checking the output matches expected patterns, without doing a full re-build and serve.

How to use

  1. Add your model file (.gguf) to model/testdata/models/
  2. Create a corresponding JSON file with the same name (e.g. mymodel.gguf -> mymodel.json) containing: 
    {
  "prompt": "your test prompt",
  "output_contains_one": ["expected substring 1", "expected substring 2"]
}
  3. Run tests with go test ./model -run TestForwardSimple

The test will load each model file found, run the specified prompt through a forward pass using greedy sampling, and verify the output contains at least one of the expected substrings.

Why this is useful

When implementing support for a new model architecture, this provides a way to quickly verify the forward pass is working correctly before integrating it into the full system. The JSON test specification makes it easy to add new test cases without modifying code.

The current example demonstrates this with a Qwen 2.5 test that verifies the model can properly handle the greeting "hi" and respond appropriately.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/9202 **Author:** [@BruceMacD](https://github.com/BruceMacD) **Created:** 2/18/2025 **Status:** ❌ Closed **Base:** `main` ← **Head:** `brucemacd/forward-test` --- ### 📝 Commits (1) - [`31d04eb`](https://github.com/ollama/ollama/commit/31d04eb7950e058c598f997c51bac1391b715ca5) model: add a test for model forward pass during implementation ### 📊 Changes **4 files changed** (+158 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `.gitignore` (+3 -0) ➕ `model/model_external_test.go` (+138 -0) ➕ `model/testdata/models/README.md` (+10 -0) ➕ `model/testdata/models/qwen2_5.json` (+7 -0) </details> ### 📄 Description This PR introduces a new testing framework for validating model forward pass behavior using JSON-specified test cases. It helps developers verify their model implementations by running simple prompts through the model and checking the output matches expected patterns, without doing a full re-build and serve. ## How to use 1. Add your model file (`.gguf`) to `model/testdata/models/` 2. Create a corresponding JSON file with the same name (e.g. `mymodel.gguf` -> `mymodel.json`) containing: ```json { "prompt": "your test prompt", "output_contains_one": ["expected substring 1", "expected substring 2"] } ``` 3. Run tests with `go test ./model -run TestForwardSimple` The test will load each model file found, run the specified prompt through a forward pass using greedy sampling, and verify the output contains at least one of the expected substrings. ## Why this is useful When implementing support for a new model architecture, this provides a way to quickly verify the forward pass is working correctly before integrating it into the full system. The JSON test specification makes it easy to add new test cases without modifying code. The current example demonstrates this with a Qwen 2.5 test that verifies the model can properly handle the greeting "hi" and respond appropriately. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-22 23:25:35 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#38760
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?