llm-multiverse/llm-multiverse
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

Pull and benchmark Qwen2.5 14B Instruct #4

New Issue
Closed
opened 2026-03-08 10:46:59 +01:00 by shahondin1624 · 1 comment
shahondin1624 commented 2026-03-08 10:46:59 +01:00
Owner
Copy Link

Description

Pull qwen2.5:14b-instruct via Ollama and run benchmarks to establish baseline performance metrics.

Acceptance Criteria

  • Model pulled successfully
  • Tokens/second measured for generation
  • VRAM usage recorded
  • Prompt evaluation speed measured
  • Results documented in a benchmark table

Blocked by

## Description Pull `qwen2.5:14b-instruct` via Ollama and run benchmarks to establish baseline performance metrics. ## Acceptance Criteria - [ ] Model pulled successfully - [ ] Tokens/second measured for generation - [ ] VRAM usage recorded - [ ] Prompt evaluation speed measured - [ ] Results documented in a benchmark table ## Blocked by - #1
shahondin1624 added this to the Phase 0: ROCm / Ollama Verification milestone 2026-03-08 10:46:59 +01:00
shahondin1624 commented 2026-03-08 21:31:11 +01:00
Author
Owner
Copy Link

Benchmark Results — qwen2.5:14b-instruct

Metric Value
VRAM usage 9.7 GB
Processor 100% GPU
Prompt eval rate 625 tok/s
Generation rate 54 tok/s
Context size 4096

Model pulled and benchmarked successfully. Nearly identical performance to qwen2.5-coder:14b — same VRAM footprint, similar generation rate. Good candidate for researcher / summarization role.

## Benchmark Results — qwen2.5:14b-instruct | Metric | Value | |---|---| | VRAM usage | 9.7 GB | | Processor | 100% GPU | | Prompt eval rate | 625 tok/s | | Generation rate | 54 tok/s | | Context size | 4096 | Model pulled and benchmarked successfully. Nearly identical performance to qwen2.5-coder:14b — same VRAM footprint, similar generation rate. Good candidate for researcher / summarization role.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
cat:agent
Agent related
 cat:cross-cutting
Cross-cutting concern
 cat:deployment
Deployment related
 cat:devops
DevOps related
 cat:performance
Performance related
 cat:security
Security related
 cat:tech-debt
Technical debt from code reviews
 lang:protobuf
Protocol Buffers
 lang:python
Python language
 lang:rust
Rust language
 priority:critical
Critical priority
 priority:high
High priority
 priority:low
Low priority
 priority:medium
Medium priority
 service:audit
Audit service
 service:memory
Memory service
 service:model-gateway
Model Gateway service
 service:orchestrator
Orchestrator service
 service:search
Search service
 service:secrets
Secrets service
 service:tool-broker
Tool Broker service
 type:bug
Bug fix
 type:documentation
Docs
 type:feature
New functionality
 type:infrastructure
Build tooling, project setup
 type:refactor
Restructuring
 type:research
Investigation, benchmarking
 type:testing
Tests
 workflow:blocked
Blocked after max retries
 workflow:manual
Requires manual intervention — blocked from auto-dev
No Label
Milestone
No items
No Milestone Phase 0: ROCm / Ollama Verification
Projects
Clear projects
No project
Assignees
Clear assignees
shahondin1624
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: llm-multiverse/llm-multiverse#4
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?