chore: import upstream snapshot with attribution
CI / benchmark (push) Has been skipped
install-script / posix-syntax (push) Successful in 6m1s
CI / build-onnx (push) Failing after 6m43s
init-smoke / dry-run (push) Failing after 15m57s
security / govulncheck (push) Has been cancelled
security / trivy-fs (push) Has been cancelled
CI / test (1.26, ubuntu-latest) (push) Has been cancelled
Scorecard supply-chain security / Scorecard analysis (push) Has been cancelled
CI / test (1.26, macos-latest) (push) Has been cancelled
CI / build-windows (push) Has been cancelled
CI / lint (push) Has been cancelled
install-script / powershell-syntax (push) Has been cancelled
install-script / install (macos-14) (push) Has been cancelled
install-script / install (ubuntu-latest) (push) Has been cancelled
CI / benchmark (push) Has been skipped
install-script / posix-syntax (push) Successful in 6m1s
CI / build-onnx (push) Failing after 6m43s
init-smoke / dry-run (push) Failing after 15m57s
security / govulncheck (push) Has been cancelled
security / trivy-fs (push) Has been cancelled
CI / test (1.26, ubuntu-latest) (push) Has been cancelled
Scorecard supply-chain security / Scorecard analysis (push) Has been cancelled
CI / test (1.26, macos-latest) (push) Has been cancelled
CI / build-windows (push) Has been cancelled
CI / lint (push) Has been cancelled
install-script / powershell-syntax (push) Has been cancelled
install-script / install (macos-14) (push) Has been cancelled
install-script / install (ubuntu-latest) (push) Has been cancelled
This commit is contained in:
@@ -0,0 +1,32 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"github.com/spf13/cobra"
|
||||
)
|
||||
|
||||
var evalCmd = &cobra.Command{
|
||||
Use: "eval",
|
||||
Short: "Retrieval-quality and token-savings benchmarks",
|
||||
Long: `Gortex evaluation harness.
|
||||
|
||||
Subcommands:
|
||||
recall — recall@1/5/20 over a curated query fixture, per ranker
|
||||
embedders — quality vs speed comparison across MiniLM-L6-v2 ONNX variants
|
||||
swebench — run the SWE-bench harness (eval/run_eval.py)
|
||||
stdbench — standardized retrieval benches (CoIR / SWE-ContextBench / ContextBench)
|
||||
tokens — GCX1 wire-format token savings vs JSON
|
||||
parity — per-language cross-file coverage vs the committed baseline
|
||||
|
||||
Typical use:
|
||||
|
||||
gortex eval recall --fixture bench/fixtures/retrieval.yaml --format markdown
|
||||
gortex eval embedders --variants fp32,qint8_arm64
|
||||
gortex eval swebench --list-configs
|
||||
gortex eval stdbench --bench coir --dataset datasets/coir/codesearchnet
|
||||
gortex eval tokens --cases bench/wire-format/cases
|
||||
`,
|
||||
}
|
||||
|
||||
func init() {
|
||||
rootCmd.AddCommand(evalCmd)
|
||||
}
|
||||
Reference in New Issue
Block a user