diff --git a/go/aio/aio.go b/go/aio/aio.go index 771d97e..fac09a5 100644 --- a/go/aio/aio.go +++ b/go/aio/aio.go @@ -27,6 +27,10 @@ import ( "github.com/google/zimtohrli/go/audio" ) +const ( + DefaultSampleRate = 48000 +) + // Fetch calls Recode if path ends with .wav, otherwise Copy. func Fetch(path string, dir string) (string, error) { if strings.ToLower(filepath.Ext(path)) == ".wav" { @@ -37,7 +41,7 @@ func Fetch(path string, dir string) (string, error) { // Load loads audio from an ffmpeg-decodable file from a path (which may be a URL). func Load(path string) (*audio.Audio, error) { - return LoadAtRate(path, 48000) + return LoadAtRate(path, DefaultSampleRate) } // LoadAtRate loads audio from an ffmpeg-decodable file from a path (which may be a URL) and returns it at the given sample rate. diff --git a/go/bin/score/score.go b/go/bin/score/score.go index 3c66d0c..4d92a53 100644 --- a/go/bin/score/score.go +++ b/go/bin/score/score.go @@ -25,6 +25,7 @@ import ( "runtime" "sort" + "github.com/google/zimtohrli/go/aio" "github.com/google/zimtohrli/go/data" "github.com/google/zimtohrli/go/goohrli" "github.com/google/zimtohrli/go/pipe" @@ -32,10 +33,6 @@ import ( "github.com/google/zimtohrli/go/worker" ) -const ( - sampleRate = 48000 -) - func main() { details := flag.String("details", "", "Glob to directories with databases to show the details of.") calculate := flag.String("calculate", "", "Glob to directories with databases to calculate metrics for.") @@ -44,7 +41,7 @@ func main() { zimtohrliScoreType := flag.String("zimtohrli_score_type", string(data.Zimtohrli), "Score type name to use when storing Zimtohrli scores in a dataset.") calculateViSQOL := flag.Bool("calculate_visqol", false, "Whether to calculate ViSQOL scores.") calculatePipeMetric := flag.String("calculate_pipe", "", "Path to a binary that serves metrics via stdin/stdout pipe. Install some of the via 'install_python_metrics.py'.") - zimtohrliParameters := goohrli.DefaultParameters(48000) + zimtohrliParameters := goohrli.DefaultParameters(aio.DefaultSampleRate) b, err := json.Marshal(zimtohrliParameters) if err != nil { log.Panic(err) @@ -55,16 +52,15 @@ func main() { report := flag.String("report", "", "Glob to directories with databases to generate a report for.") accuracy := flag.String("accuracy", "", "Glob to directories with databases to provide JND accuracy for.") mse := flag.String("mse", "", "Glob to directories with databases to provide mean-square-error when predicting MOS or JND for.") + optimizedMSE := flag.String("optimized_mse", "", "Glob to directories with databases to provide mean-square-error when predicting MOS or JND for after having optimized the MOS mapping (as in `-optimize_mapping`).") optimize := flag.String("optimize", "", "Glob to directories with databases to optimize for.") optimizeLogfile := flag.String("optimize_logfile", "", "File to write optimization events to.") - optimizeStartStep := flag.Float64("optimize_start_step", 1, "Start step for the simulated annealing.") - optimizeNumSteps := flag.Float64("optimize_num_steps", 1000, "Number of steps for the simulated annealing.") workers := flag.Int("workers", runtime.NumCPU(), "Number of concurrent workers for tasks.") failFast := flag.Bool("fail_fast", false, "Whether to panic immediately on any error.") optimizeMapping := flag.String("optimize_mapping", "", "Glob to directories with databases to optimize the MOS mapping for.") flag.Parse() - if *details == "" && *calculate == "" && *correlate == "" && *accuracy == "" && *leaderboard == "" && *report == "" && *optimize == "" && *optimizeMapping == "" && *mse == "" { + if *details == "" && *calculate == "" && *correlate == "" && *accuracy == "" && *leaderboard == "" && *report == "" && *optimize == "" && *optimizeMapping == "" && *mse == "" && *optimizedMSE == "" { flag.Usage() os.Exit(1) } @@ -72,25 +68,24 @@ func main() { if err := zimtohrliParameters.Update([]byte(*zimtohrliParametersJSON)); err != nil { log.Panic(err) } + if zimtohrliParameters.SampleRate != aio.DefaultSampleRate { + log.Fatalf("Zimtohrli sample rates != %v not supported by this tool, since it loads all data set audio at %vHz.", aio.DefaultSampleRate, aio.DefaultSampleRate) + } if *optimize != "" { bundles, err := data.OpenBundles(*optimize) if err != nil { log.Fatal(err) } - optimizeLog := func(ev data.OptimizationEvent) {} + recorder := &data.Recorder{} if *optimizeLogfile != "" { f, err := os.OpenFile(*optimizeLogfile, os.O_APPEND|os.O_WRONLY|os.O_CREATE, 0644) if err != nil { log.Fatal(err) } - optimizeLog = func(ev data.OptimizationEvent) { - b, _ := json.Marshal(ev) - f.WriteString(string(b) + "\n") - f.Sync() - } + recorder.Output = f } - if err = bundles.Optimize(*optimizeStartStep, *optimizeNumSteps, optimizeLog); err != nil { + if err = bundles.Optimize(recorder); err != nil { log.Fatal(err) } } @@ -111,7 +106,6 @@ func main() { if !reflect.DeepEqual(zimtohrliParameters, goohrli.DefaultParameters(zimtohrliParameters.SampleRate)) { log.Printf("Using %+v", zimtohrliParameters) } - zimtohrliParameters.SampleRate = sampleRate z := goohrli.New(zimtohrliParameters) return z } @@ -176,18 +170,12 @@ func main() { if err != nil { log.Fatal(err) } - for _, bundle := range bundles { - if bundle.IsJND() { - fmt.Printf("Not computing correlation for JND dataset %q\n\n", bundle.Dir) - } else { - corrTable, err := bundle.Correlate() - if err != nil { - log.Fatal(err) - } - fmt.Printf("## %v\n\n", bundle.Dir) - fmt.Println(corrTable) - } + corrTable, err := bundles.Correlate() + if err != nil { + log.Fatal(err) } + fmt.Printf("## %v\n\n", *correlate) + fmt.Println(corrTable) } if *accuracy != "" { @@ -195,18 +183,26 @@ func main() { if err != nil { log.Fatal(err) } - for _, bundle := range bundles { - if bundle.IsJND() { - accuracy, err := bundle.JNDAccuracy() - if err != nil { - log.Fatal(err) - } - fmt.Printf("## %v\n", bundle.Dir) - fmt.Println(accuracy) - } else { - fmt.Printf("Not computing accuracy for non-JND dataset %q\n\n", bundle.Dir) - } + result, err := bundles.JNDAccuracy() + if err != nil { + log.Fatal(err) + } + fmt.Printf("## %v\n\n", *accuracy) + fmt.Println(result) + } + + if *optimizedMSE != "" { + bundles, err := data.OpenBundles(*optimizedMSE) + if err != nil { + log.Fatal(err) + } + res, err := bundles.OptimizedZimtohrliMSE() + if err != nil { + log.Fatal(err) } + fmt.Printf("## %v\n\n", *optimizedMSE) + fmt.Printf("Error for MOS datasets is `human-MOS - Zimtohrli-predicted-MOS`. Error for JND datasets is `distance from correct side of threshold`.\n\n") + fmt.Printf("MSE after optimizing mapping: %.15f\n\n", res) } if *mse != "" { @@ -214,15 +210,14 @@ func main() { if err != nil { log.Fatal(err) } - for _, bundle := range bundles { - z := makeZimtohrli() - mse, err := bundle.ZimtohrliMSE(z) - if err != nil { - log.Fatal(err) - } - fmt.Printf("## %v\n", bundle.Dir) - fmt.Printf("MSE: %.15f\n\n", mse) + z := makeZimtohrli() + res, err := bundles.ZimtohrliMSE(z, true) + if err != nil { + log.Fatal(err) } + fmt.Printf("## %v\n\n", *mse) + fmt.Print("Error for MOS datasets is `human-MOS - Zimtohrli-predicted-MOS`. Error for JND datasets is `distance from correct side of threshold`.\n\n") + fmt.Printf("MSE: %.15f\n\n", res) } if *report != "" { diff --git a/go/data/study.go b/go/data/study.go index 39f5c40..a1b2cdc 100644 --- a/go/data/study.go +++ b/go/data/study.go @@ -23,7 +23,6 @@ import ( "io" "log" "math" - "math/rand" "os" "os/exec" "path/filepath" @@ -96,10 +95,17 @@ type Study struct { // ReferenceBundle is a plain data type containing a bunch of references, typicall the content of a study. type ReferenceBundle struct { - Dir string - References []*Reference - ScoreTypes map[ScoreType]int + // Dir is the directory of the source study. + Dir string + // References are all the reference sounds of the bundle. + References []*Reference + // ScoreTypes are the number of scores of each type in the bundle. + ScoreTypes map[ScoreType]int + // ScoreTypeLimits are the upper/lower limits of each score type in the bundle. ScoreTypeLimits map[ScoreType][2]*float64 + + // mosScaler returns the MOS score provided scaled to 1-5. Useful for datasets where the MOS is scaled up to 0-100. + mosScaler func(float64) float64 } // ReferenceBundles is a slice of ReferenceBundle. @@ -111,6 +117,24 @@ func (r *ReferenceBundle) IsJND() bool { return res } +// ScaledMOS returns the MOS score scaled to 1-5. +func (r *ReferenceBundle) ScaledMOS(mos float64) (float64, error) { + if r.mosScaler == nil { + if math.Abs(*r.ScoreTypeLimits[MOS][0]-1) < 0.2 && math.Abs(*r.ScoreTypeLimits[MOS][1]-5) < 0.2 { + r.mosScaler = func(mos float64) float64 { + return mos + } + } else if math.Abs(*r.ScoreTypeLimits[MOS][0]) < 0.2 && math.Abs(*r.ScoreTypeLimits[MOS][1]-100) < 0.2 { + r.mosScaler = func(mos float64) float64 { + return 1 + 0.04*mos + } + } else { + return 0, fmt.Errorf("minimum MOS %v and maximum MOS %v are confusing", *r.ScoreTypeLimits[MOS][0], *r.ScoreTypeLimits[MOS][1]) + } + } + return r.mosScaler(mos), nil +} + // SortedTypes returns the score types of a bundle, alphabetically ordered. func (r *ReferenceBundle) SortedTypes() ScoreTypes { sorted := ScoreTypes{} @@ -237,10 +261,25 @@ func (c CorrelationTable) String() string { func (r *ReferenceBundle) Correlation(typeA, typeB ScoreType) (float64, error) { scoresA := []float64{} scoresB := []float64{} + appender := func(scores *[]float64, typ ScoreType, dist *Distortion) error { + score := dist.Scores[typ] + if typ == MOS { + var err error + if score, err = r.ScaledMOS(score); err != nil { + return err + } + } + *scores = append(*scores, score) + return nil + } for _, ref := range r.References { for _, dist := range ref.Distortions { - scoresA = append(scoresA, dist.Scores[typeA]) - scoresB = append(scoresB, dist.Scores[typeB]) + if err := appender(&scoresA, typeA, dist); err != nil { + return 0, err + } + if err := appender(&scoresB, typeB, dist); err != nil { + return 0, err + } } } if len(scoresA) != len(scoresB) { @@ -274,6 +313,22 @@ func (r *ReferenceBundle) Correlate() (CorrelationTable, error) { return result, nil } +// Correlate returns a table of all scores in the bundles Spearman correlated to each other. +func (r ReferenceBundles) Correlate() (CorrelationTable, error) { + merged := &ReferenceBundle{ + ScoreTypes: map[ScoreType]int{}, + ScoreTypeLimits: map[ScoreType][2]*float64{}, + } + for _, bundle := range r { + if !bundle.IsJND() { + for _, ref := range bundle.References { + merged.Add(ref) + } + } + } + return merged.Correlate() +} + // JNDAccuracyScore contains the accuracy for a metric when used to predict audible differences, and the threshold when that accuracy was achieved. type JNDAccuracyScore struct { ScoreType ScoreType @@ -399,6 +454,51 @@ func (r *ReferenceBundle) JNDAccuracy() (JNDAccuracyScores, error) { return result, nil } +// JNDAccuracy returns the accuracy of each score type when used to predict audible differences. +func (r ReferenceBundles) JNDAccuracy() (JNDAccuracyScores, error) { + merged := &ReferenceBundle{ + ScoreTypes: map[ScoreType]int{}, + ScoreTypeLimits: map[ScoreType][2]*float64{}, + } + for _, bundle := range r { + if bundle.IsJND() { + for _, ref := range bundle.References { + merged.Add(ref) + } + } + } + return merged.JNDAccuracy() +} + +// OptimizedZimtohrliMSE optimizes the MOS mapping and returns the ZimtohrliMSE using the optimized mapping. +func (r ReferenceBundles) OptimizedZimtohrliMSE() (float64, error) { + optResult, err := r.OptimizeMapping() + if err != nil { + return 0, err + } + params := goohrli.DefaultParameters(aio.DefaultSampleRate) + copy(params.MOSMapperParams[:], optResult.ParamsAfter) + z := goohrli.New(params) + return r.ZimtohrliMSE(z, true) +} + +// ZimtohrliMSE returns the mean square of the ZimtohrliMSE of the bundles. +func (r ReferenceBundles) ZimtohrliMSE(z *goohrli.Goohrli, includeJND bool) (float64, error) { + sumOfSquares := 0.0 + count := 0 + for _, bundle := range r { + if includeJND || !bundle.IsJND() { + mse, err := bundle.ZimtohrliMSE(z) + if err != nil { + return 0, err + } + sumOfSquares += mse * mse + count += 1 + } + } + return sumOfSquares / float64(count), nil +} + // ZimtohrliMSE returns the precision when predicting the MOS score or JND difference. func (r *ReferenceBundle) ZimtohrliMSE(z *goohrli.Goohrli) (float64, error) { if r.IsJND() { @@ -437,19 +537,6 @@ func (r *ReferenceBundle) ZimtohrliMSE(z *goohrli.Goohrli) (float64, error) { } return sumOfSquares / float64(count), nil } else { - var mosScaler func(mos float64) float64 - if math.Abs(*r.ScoreTypeLimits[MOS][0]-1) < 0.2 && math.Abs(*r.ScoreTypeLimits[MOS][1]-5) < 0.2 { - mosScaler = func(mos float64) float64 { - return mos - } - } else if math.Abs(*r.ScoreTypeLimits[MOS][0]) < 0.2 && math.Abs(*r.ScoreTypeLimits[MOS][1]-100) < 0.2 { - mosScaler = func(mos float64) float64 { - return 1 + 0.04*mos - } - } else { - return 0, fmt.Errorf("minimum MOS %v and maximum MOS %v are confusing", *r.ScoreTypeLimits[MOS][0], *r.ScoreTypeLimits[MOS][1]) - } - sumOfSquares := 0.0 count := 0 for _, ref := range r.References { @@ -462,7 +549,11 @@ func (r *ReferenceBundle) ZimtohrliMSE(z *goohrli.Goohrli) (float64, error) { if !found { return 0, fmt.Errorf("%+v doesn't have a Zimtohrli score", ref) } - delta := mosScaler(mos) - z.MOSFromZimtohrli(zimt) + scaledMOS, err := r.ScaledMOS(mos) + if err != nil { + return 0, err + } + delta := scaledMOS - z.MOSFromZimtohrli(zimt) sumOfSquares += delta * delta count++ } @@ -486,11 +577,8 @@ func (s Studies) ToBundles() (ReferenceBundles, error) { return result, nil } -// CalculateZimtohrliMSE returns the mean-squared-error for the Zimtohrli score -// in the bundles. For JDN bundles this means 1 - accuracy, for the MOS bundles it means -// 1 - Spearman correlation. +// CalculateZimtohrliMSE calculates Zimtohrli scores for all examples in the bundles, optimizes the MOS mapping, and returns the resulting MSE. func (r ReferenceBundles) CalculateZimtohrliMSE(z *goohrli.Goohrli) (float64, error) { - sumOfSquares := 0.0 for _, bundle := range r { bar := progress.New(fmt.Sprintf("Calculating for %v", filepath.Base(bundle.Dir))) pool := &worker.Pool[any]{ @@ -500,107 +588,9 @@ func (r ReferenceBundles) CalculateZimtohrliMSE(z *goohrli.Goohrli) (float64, er if err := bundle.Calculate(map[ScoreType]Measurement{Zimtohrli: z.NormalizedAudioDistance}, pool, true); err != nil { return 0, err } - if bundle.IsJND() { - accuracy, _, err := bundle.JNDAccuracyAndThreshold(Zimtohrli) - if err != nil { - return 0, err - } - e := (1 - accuracy) - sumOfSquares += e * e - - } else { - correlation, err := bundle.Correlation(Zimtohrli, MOS) - if err != nil { - return 0, err - } - e := (1 - correlation) - sumOfSquares += e * e - } bar.Finish() } - return sumOfSquares / float64(len(r)), nil -} - -func mutateFloat(f, min, max float64, rng *rand.Rand, temp float64) float64 { - r := math.Sqrt(temp) * rng.NormFloat64() * 0.2 * (max - min) - if f == min || r > 0 { - f += math.Abs(r) - } else if f == max || r < 0 { - f -= math.Abs(r) - } else if r == 0 { - return f - } - if f < min { - f = min - } - if f > max { - f = max - } - return f -} - -func mutateInt(i, min, max int, rng *rand.Rand, temp float64) int { - if float64(i)*temp < 1 { - i += (rng.Int() % 3) - 1 - if i < min { - i = min - } - if i > max { - i = max - } - return i - } - return int(mutateFloat(float64(i), float64(min), float64(max), rng, temp)) -} - -const sampleRate = 48000 - -func mutate(z *goohrli.Goohrli, rng *rand.Rand, temp float64) *goohrli.Goohrli { - params := z.Parameters() - params.PerceptualSampleRate = mutateFloat(params.PerceptualSampleRate, 50, 150, rng, temp) - params.FrequencyResolution = mutateFloat(params.FrequencyResolution, 1, 15, rng, temp) - params.NSIMChannelWindow = mutateInt(params.NSIMChannelWindow, 3, 64, rng, temp) - params.NSIMStepWindow = mutateInt(params.NSIMStepWindow, 3, 64, rng, temp) - result := goohrli.New(params) - return result -} - -// References returns the sum of the number of references in all the bundles. -func (r ReferenceBundles) References() int { - res := 0 - for _, bundle := range r { - res += len(bundle.References) - } - return res -} - -// Split will split the bundle randomly in two parts, at the split provided. -func (r ReferenceBundles) Split(rng *rand.Rand, split float64) (ReferenceBundles, ReferenceBundles) { - left := ReferenceBundles{} - right := ReferenceBundles{} - for _, bundle := range r { - newLeft := &ReferenceBundle{ - Dir: bundle.Dir, - References: nil, - ScoreTypes: map[ScoreType]int{}, - } - left = append(left, newLeft) - newRight := &ReferenceBundle{ - Dir: bundle.Dir, - References: nil, - ScoreTypes: map[ScoreType]int{}, - } - right = append(right, newRight) - numLeft := int(split * float64(len(bundle.References))) - indices := rng.Perm(len(bundle.References)) - for _, index := range indices[:numLeft] { - newLeft.Add(bundle.References[index]) - } - for _, index := range indices[numLeft:] { - newRight.Add(bundle.References[index]) - } - } - return left, right + return r.OptimizedZimtohrliMSE() } // MappingOptimizationResult contains the results of optimizing the MOS mapping. @@ -613,28 +603,20 @@ type MappingOptimizationResult struct { // OptimizeMOSMapping optimizes the MOS mapping parameters. func (r ReferenceBundles) OptimizeMapping() (*MappingOptimizationResult, error) { - z := goohrli.New(goohrli.DefaultParameters(48000)) + startParams := goohrli.DefaultParameters(aio.DefaultSampleRate) errors := []error{} p := optimize.Problem{ Func: func(x []float64) float64 { - params := z.Parameters() + params := startParams for index := range params.MOSMapperParams { params.MOSMapperParams[index] = math.Abs(x[index]) } - z.Set(params) - sum := 0.0 - count := 0 - for _, bundle := range r { - if !bundle.IsJND() { - mse, err := bundle.ZimtohrliMSE(z) - if err != nil { - errors = append(errors, err) - } - sum += mse - count += 1 - } + z := goohrli.New(params) + result, err := r.ZimtohrliMSE(z, false) + if err != nil { + errors = append(errors, err) } - return sum / float64(count) + return result }, Status: func() (optimize.Status, error) { if len(errors) > 0 { @@ -643,12 +625,11 @@ func (r ReferenceBundles) OptimizeMapping() (*MappingOptimizationResult, error) return optimize.NotTerminated, nil }, } - startParams := z.Parameters().MOSMapperParams result := &MappingOptimizationResult{ - ParamsBefore: startParams[:], - MSEBefore: p.Func(startParams[:]), + ParamsBefore: startParams.MOSMapperParams[:], + MSEBefore: p.Func(startParams.MOSMapperParams[:]), } - optResult, err := optimize.Minimize(p, startParams[:], nil, nil) + optResult, err := optimize.Minimize(p, startParams.MOSMapperParams[:], &optimize.Settings{Concurrent: runtime.NumCPU()}, nil) if err != nil { return nil, err } @@ -660,50 +641,90 @@ func (r ReferenceBundles) OptimizeMapping() (*MappingOptimizationResult, error) return result, nil } -// OptimizationEvent is a step in the optimization process. -type OptimizationEvent struct { +// OptimizeEvent is a step in the optimization process. +type OptimizeEvent struct { Parameters goohrli.Parameters Step int - Loss float64 - Temp float64 + MSE float64 +} + +// Recorder logs optimization progress. +type Recorder struct { + Output *os.File + + startParameters goohrli.Parameters } -// Optimize will use simulated annealing to optimize a Zimtohrli metric for predicting -// these bundles. -func (r ReferenceBundles) Optimize(startStep, numSteps float64, logger func(OptimizationEvent)) error { - z := goohrli.New(goohrli.DefaultParameters(sampleRate)) - loss, err := r.CalculateZimtohrliMSE(z) +func (r *Recorder) Init() error { + return nil +} + +func (r *Recorder) paramsToX(p goohrli.Parameters) []float64 { + return []float64{p.FrequencyResolution / r.startParameters.FrequencyResolution} +} + +func (r *Recorder) xToParams(x []float64) goohrli.Parameters { + cpy := r.startParameters + cpy.FrequencyResolution *= x[0] + return cpy +} + +func (r *Recorder) Record(loc *optimize.Location, op optimize.Operation, stats *optimize.Stats) error { + params := r.xToParams(loc.X) + switch op { + case optimize.InitIteration: + log.Printf("Initialized solution %+v with MSE %v", params, loc.F) + case optimize.MajorIteration: + log.Printf("%v iterations, candidate solution %+v with MSE %v", stats.MajorIterations, params, loc.F) + case optimize.FuncEvaluation: + log.Printf("%v iterations, evaluated at %+v with MSE %v", stats.MajorIterations, params, loc.F) + case optimize.MethodDone: + log.Printf("Solution %+v found with MSE %v", loc.F, loc.X) + } + if r.Output == nil { + return nil + } + ev := OptimizeEvent{ + Parameters: params, + Step: stats.MajorIterations, + MSE: loc.F, + } + b, err := json.Marshal(ev) if err != nil { return err } - logger(OptimizationEvent{Parameters: z.Parameters(), Step: 0, Loss: loss, Temp: 1}) - log.Printf("Created initial solution %v with loss %.2f", z, loss) - for step := startStep; step < numSteps; step++ { - rng := rand.New(rand.NewSource(int64(step))) - temp := 1.0 - (step+1)/numSteps - newZ := mutate(z, rng, temp) - log.Printf("Created new solution %+v", newZ) - newLoss, err := r.CalculateZimtohrliMSE(newZ) - if err != nil { - return err - } - log.Printf("Step %v, temp %v, old loss %.2f, new loss %.2f", step, temp, loss, newLoss) - logger(OptimizationEvent{Parameters: newZ.Parameters(), Step: int(step), Loss: newLoss, Temp: temp}) - if newLoss < loss { - z = newZ - loss = newLoss - log.Print("*** Accepting better solution") - } else { - acceptanceProb := math.Exp(-(newLoss - loss) / temp) - dice := rng.Float64() - if dice < acceptanceProb { - z = newZ - loss = newLoss - log.Printf("*** Accepting poorer solution due to acceptanceProb=%.2f > dice=%.2f", acceptanceProb, dice) - } else { - log.Print("Discarding poorer solution") + if _, err := r.Output.WriteString(string(b) + "\n"); err != nil { + return err + } + return r.Output.Sync() +} + +// Optimize will use optimize a Zimtohrli metric for predicting these bundles. +func (r ReferenceBundles) Optimize(recorder *Recorder) error { + recorder.startParameters = goohrli.DefaultParameters(aio.DefaultSampleRate) + errors := []error{} + p := optimize.Problem{ + Func: func(x []float64) float64 { + z := goohrli.New(recorder.xToParams(x)) + mse, err := r.CalculateZimtohrliMSE(z) + if err != nil { + errors = append(errors, err) } - } + return mse + }, + Status: func() (optimize.Status, error) { + if len(errors) > 0 { + return optimize.Failure, fmt.Errorf("%+v", errors) + } + return optimize.NotTerminated, nil + }, + } + optResult, err := optimize.Minimize(p, recorder.paramsToX(recorder.startParameters), &optimize.Settings{Recorder: recorder}, nil) + if err != nil { + return err + } + if err := optResult.Status.Err(); err != nil { + return err } return nil } diff --git a/python/mos_mapping.ipynb b/python/mos_mapping.ipynb index 9ab15fe..8aae1dd 100644 --- a/python/mos_mapping.ipynb +++ b/python/mos_mapping.ipynb @@ -21,7 +21,7 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 62, "metadata": { "id": "gsCcH5KtJ2x9" }, @@ -36,7 +36,7 @@ }, { "cell_type": "code", - "execution_count": 39, + "execution_count": 63, "metadata": {}, "outputs": [ { @@ -112,430 +112,30 @@ }, { "cell_type": "code", - "execution_count": 42, + "execution_count": 73, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "params=array([1., 1., 1.]) result=3.434709402324488\n", - "params=array([1.00000001, 1. , 1. ]) result=3.434709402324488\n", - "params=array([1. , 1.00000001, 1. ]) result=3.4347094125550726\n", - "params=array([1. , 1. , 1.00000001]) result=3.4347093778098983\n", - "params=array([1. , 0.6110146 , 1.93208924]) result=1.8939908848446885\n", - "params=array([1.00000001, 0.6110146 , 1.93208924]) result=1.8939908848446878\n", - "params=array([1. , 0.61101461, 1.93208924]) result=1.8939908976347193\n", - "params=array([1. , 0.6110146 , 1.93208926]) result=1.8939908706072108\n", - "params=array([ 1.0000001 , -1.12345646, 3.63128219]) result=1.193248803218663\n", - "params=array([ 1.00000011, -1.12345646, 3.63128219]) result=1.1932488032186632\n", - "params=array([ 1.0000001 , -1.12345645, 3.63128219]) result=1.1932487973174661\n", - "params=array([ 1.0000001 , -1.12345646, 3.6312822 ]) result=1.1932487981486903\n", - "params=array([ 1.0000001 , -0.90849519, 3.89278154]) result=1.0388598245721532\n", - "params=array([ 1.00000012, -0.90849519, 3.89278154]) result=1.0388598245721532\n", - "params=array([ 1.0000001 , -0.90849518, 3.89278154]) result=1.0388598194919634\n", - "params=array([ 1.0000001 , -0.90849519, 3.89278156]) result=1.0388598210821653\n", - "params=array([ 1.00000014, -0.28240301, 4.31298199]) result=0.8243722141104614\n", - "params=array([ 1.00000015, -0.28240301, 4.31298199]) result=0.8243722141104612\n", - "params=array([ 1.00000014, -0.28240299, 4.31298199]) result=0.8243722120539925\n", - "params=array([ 1.00000014, -0.28240301, 4.312982 ]) result=0.8243722137435218\n", - "params=array([1.00000016, 0.10591622, 4.39865415]) result=0.8086392778342267\n", - "params=array([1.00000018, 0.10591622, 4.39865415]) result=0.8086392778342267\n", - "params=array([1.00000016, 0.10591624, 4.39865415]) result=0.8086392783565501\n", - "params=array([1.00000016, 0.10591622, 4.39865417]) result=0.8086392782262002\n", - "params=array([1.00000016, 0.02684716, 4.35582085]) result=0.8059468035381968\n", - "params=array([1.00000017, 0.02684716, 4.35582085]) result=0.8059468035381968\n", - "params=array([1.00000016, 0.02684718, 4.35582085]) result=0.805946803462382\n", - "params=array([1.00000016, 0.02684716, 4.35582087]) result=0.8059468041559786\n", - "params=array([ 1.00000015, -0.05638057, 4.21442093]) result=0.8022680530556958\n", - "params=array([ 1.00000016, -0.05638057, 4.21442093]) result=0.8022680530556958\n", - "params=array([ 1.00000015, -0.05638056, 4.21442093]) result=0.802268052295488\n", - "params=array([ 1.00000015, -0.05638057, 4.21442095]) result=0.8022680533492818\n", - "params=array([ 1.00000015, -0.01472949, 4.19360523]) result=0.8001001682313968\n", - "params=array([ 1.00000016, -0.01472949, 4.19360523]) result=0.8001001682313968\n", - "params=array([ 1.00000015, -0.01472948, 4.19360523]) result=0.8001001678068568\n", - "params=array([ 1.00000015, -0.01472949, 4.19360525]) result=0.8001001686578644\n", - "params=array([1.00000014, 0.04260846, 4.09767338]) result=0.7996542476058521\n", - "params=array([1.00000015, 0.04260846, 4.09767338]) result=0.7996542476058521\n", - "params=array([1.00000014, 0.04260848, 4.09767338]) result=0.7996542487075058\n", - "params=array([1.00000014, 0.04260846, 4.09767339]) result=0.7996542477581412\n", - "params=array([1.00000014, 0.03118503, 4.09217264]) result=0.7987804419199129\n", - "params=array([1.00000015, 0.03118503, 4.09217264]) result=0.798780441919913\n", - "params=array([1.00000014, 0.03118505, 4.09217264]) result=0.7987804429323027\n", - "params=array([1.00000014, 0.03118503, 4.09217266]) result=0.7987804421105288\n", - "params=array([ 1.00000014, -0.01450868, 4.0701697 ]) result=0.7973843997555085\n", - "params=array([ 1.00000015, -0.01450868, 4.0701697 ]) result=0.7973843997555086\n", - "params=array([ 1.00000014, -0.01450867, 4.0701697 ]) result=0.7973843988194768\n", - "params=array([ 1.00000014, -0.01450868, 4.07016972]) result=0.797384399978586\n", - "params=array([1.00000013, 0.00725317, 4.02801307]) result=0.796332443246752\n", - "params=array([1.00000014, 0.00725317, 4.02801307]) result=0.796332443246752\n", - "params=array([1.00000013, 0.00725318, 4.02801307]) result=0.7963324442898436\n", - "params=array([1.00000013, 0.00725317, 4.02801308]) result=0.7963324434265853\n", - "params=array([ 1.00000012e+00, -2.05922155e-03, 3.95142667e+00]) result=0.7952967094748329\n", - "params=array([ 1.00000013e+00, -2.05922155e-03, 3.95142667e+00]) result=0.7952967094748329\n", - "params=array([ 1.00000012e+00, -2.05920665e-03, 3.95142667e+00]) result=0.7952967081433059\n", - "params=array([ 1.00000012e+00, -2.05922155e-03, 3.95142668e+00]) result=0.7952967095358355\n", - "params=array([1.00000011, 0.00470432, 3.88130072]) result=0.7956287028225006\n", - "params=array([1.00000013, 0.00470432, 3.88130072]) result=0.7956287028225006\n", - "params=array([1.00000011, 0.00470434, 3.88130072]) result=0.7956287045074268\n", - "params=array([1.00000011, 0.00470432, 3.88130073]) result=0.795628702732547\n", - "params=array([1.00000012e+00, 6.75084684e-05, 3.92937624e+00]) result=0.7950477465548658\n", - "params=array([1.00000013e+00, 6.75084684e-05, 3.92937624e+00]) result=0.7950477465548659\n", - "params=array([1.00000012e+00, 6.75233695e-05, 3.92937624e+00]) result=0.7950477479667526\n", - "params=array([1.00000012e+00, 6.75084684e-05, 3.92937626e+00]) result=0.7950477465824624\n", - "params=array([ 1.00000011e+00, -1.18976882e-03, 3.89779376e+00]) result=0.795172892099672\n", - "params=array([ 1.00000012e+00, -1.18976882e-03, 3.89779376e+00]) result=0.7951728920996721\n", - "params=array([ 1.00000011e+00, -1.18975392e-03, 3.89779376e+00]) result=0.7951728905288844\n", - "params=array([ 1.00000011e+00, -1.18976882e-03, 3.89779377e+00]) result=0.7951728920596406\n", - "params=array([ 1.00000012e+00, -1.79999898e-04, 3.92315889e+00]) result=0.7950496425612137\n", - "params=array([ 1.00000013e+00, -1.79999898e-04, 3.92315889e+00]) result=0.7950496425612137\n", - "params=array([ 1.00000012e+00, -1.79984997e-04, 3.92315889e+00]) result=0.795049641119352\n", - "params=array([ 1.00000012e+00, -1.79999898e-04, 3.92315891e+00]) result=0.7950496425761503\n", - "params=array([ 1.00000012e+00, -2.68177438e-05, 3.92700679e+00]) result=0.7950398406396849\n", - "params=array([ 1.00000013e+00, -2.68177438e-05, 3.92700679e+00]) result=0.7950398406396849\n", - "params=array([ 1.00000012e+00, -2.68028426e-05, 3.92700679e+00]) result=0.795039839217266\n", - "params=array([ 1.00000012e+00, -2.68177438e-05, 3.92700681e+00]) result=0.7950398406628517\n", - "params=array([ 1.00000011e+00, -4.52345981e-06, 3.91810358e+00]) result=0.7950290118406604\n", - "params=array([ 1.00000013e+00, -4.52345981e-06, 3.91810358e+00]) result=0.7950290118406604\n", - "params=array([ 1.00000011e+00, -4.50855864e-06, 3.91810358e+00]) result=0.7950290103771738\n", - "params=array([ 1.00000011e+00, -4.52345981e-06, 3.91810359e+00]) result=0.7950290118464923\n", - "params=array([1.00000011e+00, 6.34342414e-05, 3.91306514e+00]) result=0.7950345971845392\n", - "params=array([1.00000012e+00, 6.34342414e-05, 3.91306514e+00]) result=0.7950345971845391\n", - "params=array([1.00000011e+00, 6.34491426e-05, 3.91306514e+00]) result=0.7950345986720988\n", - "params=array([1.00000011e+00, 6.34342414e-05, 3.91306516e+00]) result=0.7950345971801666\n", - "params=array([1.00000011e+00, 8.56397602e-06, 3.91713327e+00]) result=0.7950290922892727\n", - "params=array([1.00000013e+00, 8.56397602e-06, 3.91713327e+00]) result=0.7950290922892727\n", - "params=array([1.00000011e+00, 8.57887718e-06, 3.91713327e+00]) result=0.7950290937573142\n", - "params=array([1.00000011e+00, 8.56397602e-06, 3.91713328e+00]) result=0.7950290922931769\n", - "params=array([1.00000011e+00, 7.75985023e-07, 3.91771067e+00]) result=0.795028499751508\n", - "params=array([1.00000013e+00, 7.75985023e-07, 3.91771067e+00]) result=0.795028499751508\n", - "params=array([1.00000011e+00, 7.90886184e-07, 3.91771067e+00]) result=0.7950285012167823\n", - "params=array([1.00000011e+00, 7.75985023e-07, 3.91771069e+00]) result=0.7950284997565847\n", - "params=array([ 1.00000011e+00, -1.56964360e-06, 3.91535146e+00]) result=0.7950281422655635\n", - "params=array([ 1.00000013e+00, -1.56964360e-06, 3.91535146e+00]) result=0.7950281422655635\n", - "params=array([ 1.00000011e+00, -1.55474244e-06, 3.91535146e+00]) result=0.7950281407893044\n", - "params=array([ 1.00000011e+00, -1.56964360e-06, 3.91535147e+00]) result=0.7950281422659911\n", - "params=array([1.00000011e+00, 2.84269852e-07, 3.91470634e+00]) result=0.7950280237005174\n", - "params=array([1.00000013e+00, 2.84269852e-07, 3.91470634e+00]) result=0.7950280237005174\n", - "params=array([1.00000011e+00, 2.99171013e-07, 3.91470634e+00]) result=0.795028025179767\n", - "params=array([1.00000011e+00, 2.84269852e-07, 3.91470635e+00]) result=0.7950280236996785\n", - "params=array([ 1.00000011e+00, -6.82570573e-07, 3.91504278e+00]) result=0.7950280517503074\n", - "params=array([ 1.00000013e+00, -6.82570573e-07, 3.91504278e+00]) result=0.7950280517503072\n", - "params=array([ 1.00000011e+00, -6.67669412e-07, 3.91504278e+00]) result=0.7950280502726205\n", - "params=array([ 1.00000011e+00, -6.82570573e-07, 3.91504279e+00]) result=0.7950280517501304\n", - "params=array([1.00000013e+00, 1.19546023e-06, 3.91475703e+00]) result=0.795028111457819\n", - "params=array([1.00000014e+00, 1.19546023e-06, 3.91475703e+00]) result=0.795028111457819\n", - "params=array([1.00000013e+00, 1.21036139e-06, 3.91475703e+00]) result=0.7950281129368422\n", - "params=array([1.00000013e+00, 1.19546023e-06, 3.91475704e+00]) result=0.795028111457076\n", - "params=array([ 1.00000012e+00, -1.05038956e-07, 3.91495490e+00]) result=0.7950279960188825\n", - "params=array([ 1.00000013e+00, -1.05038956e-07, 3.91495490e+00]) result=0.7950279960188825\n", - "params=array([ 1.00000012e+00, -9.01377951e-08, 3.91495490e+00]) result=0.7950279945407924\n", - "params=array([ 1.00000012e+00, -1.05038956e-07, 3.91495492e+00]) result=0.7950279960185347\n", - "params=array([1.00000013e+00, 7.53290505e-07, 3.91482430e+00]) result=0.7950280645239786\n", - "params=array([1.00000014e+00, 7.53290505e-07, 3.91482430e+00]) result=0.7950280645239786\n", - "params=array([1.00000013e+00, 7.68191666e-07, 3.91482430e+00]) result=0.7950280660026838\n", - "params=array([1.00000013e+00, 7.53290505e-07, 3.91482432e+00]) result=0.7950280645233703\n", - "params=array([1.00000012e+00, 3.75050303e-08, 3.91493321e+00]) result=0.7950279898568355\n", - "params=array([1.00000013e+00, 3.75050303e-08, 3.91493321e+00]) result=0.7950279898568354\n", - "params=array([1.00000012e+00, 5.24061915e-08, 3.91493321e+00]) result=0.7950279913350262\n", - "params=array([1.00000012e+00, 3.75050303e-08, 3.91493323e+00]) result=0.7950279898564454\n", - "params=array([ 1.00000012e+00, -3.64705502e-08, 3.91494447e+00]) result=0.7950279894678463\n", - "params=array([ 1.00000013e+00, -3.64705502e-08, 3.91494447e+00]) result=0.7950279894678465\n", - "params=array([ 1.00000012e+00, -2.15693890e-08, 3.91494447e+00]) result=0.7950279879897084\n", - "params=array([ 1.00000012e+00, -3.64705502e-08, 3.91494448e+00]) result=0.7950279894674784\n", - "params=array([ 1.00000012e+00, -9.24653272e-10, 3.91493906e+00]) result=0.7950279860772803\n", - "params=array([ 1.00000013e+00, -9.24653272e-10, 3.91493906e+00]) result=0.7950279860772805\n", - "params=array([1.00000012e+00, 1.39765079e-08, 3.91493906e+00]) result=0.7950279873719961\n", - "params=array([ 1.00000012e+00, -9.24653272e-10, 3.91493908e+00]) result=0.7950279860769017\n", - "params=array([ 1.00000012e+00, -1.82501350e-08, 3.91494170e+00]) result=0.7950279877294129\n", - "params=array([ 1.00000013e+00, -1.82501350e-08, 3.91494170e+00]) result=0.7950279877294129\n", - "params=array([ 1.00000012e+00, -3.34897385e-09, 3.91494170e+00]) result=0.7950279862512623\n", - "params=array([ 1.00000012e+00, -1.82501350e-08, 3.91494171e+00]) result=0.7950279877290394\n", - "params=array([ 1.00000012e+00, -3.27625592e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625592e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249053e-08, 3.91493942e+00]) result=0.7950279871296406\n", - "params=array([ 1.00000012e+00, -3.27625592e-09, 3.91493943e+00]) result=0.7950279863010957\n", - "params=array([ 1.00000012e+00, -2.53458755e-07, 3.91506263e+00]) result=0.7950280089897778\n", - "params=array([ 1.00000014e+00, -2.53458755e-07, 3.91506263e+00]) result=0.7950280089897777\n", - "params=array([ 1.00000012e+00, -2.38557594e-07, 3.91506263e+00]) result=0.7950280075121877\n", - "params=array([ 1.00000012e+00, -2.53458755e-07, 3.91506265e+00]) result=0.7950280089896419\n", - "params=array([ 1.00000012e+00, -3.30592362e-08, 3.91495409e+00]) result=0.7950279888980702\n", - "params=array([ 1.00000013e+00, -3.30592362e-08, 3.91495409e+00]) result=0.7950279888980702\n", - "params=array([ 1.00000012e+00, -1.81580750e-08, 3.91495409e+00]) result=0.7950279874199769\n", - "params=array([ 1.00000012e+00, -3.30592362e-08, 3.91495410e+00]) result=0.7950279888977212\n", - "params=array([ 1.00000012e+00, -6.87053952e-09, 3.91494119e+00]) result=0.7950279866133408\n", - "params=array([ 1.00000013e+00, -6.87053952e-09, 3.91494119e+00]) result=0.795027986613341\n", - "params=array([1.00000012e+00, 8.03062167e-09, 3.91494119e+00]) result=0.7950279867284179\n", - "params=array([ 1.00000012e+00, -6.87053952e-09, 3.91494120e+00]) result=0.7950279866129665\n", - "params=array([ 1.00000012e+00, -3.59689119e-09, 3.91493958e+00]) result=0.7950279863292775\n", - "params=array([ 1.00000013e+00, -3.59689119e-09, 3.91493958e+00]) result=0.7950279863292774\n", - "params=array([1.00000012e+00, 1.13042700e-08, 3.91493958e+00]) result=0.7950279870938315\n", - "params=array([ 1.00000012e+00, -3.59689119e-09, 3.91493959e+00]) result=0.7950279863289\n", - "params=array([ 1.00000012e+00, -3.30203897e-09, 3.91493943e+00]) result=0.7950279863037093\n", - "params=array([ 1.00000013e+00, -3.30203897e-09, 3.91493943e+00]) result=0.7950279863037091\n", - "params=array([1.00000012e+00, 1.15991222e-08, 3.91493943e+00]) result=0.795027987126761\n", - "params=array([ 1.00000012e+00, -3.30203897e-09, 3.91493945e+00]) result=0.7950279863033314\n", - "params=array([ 1.00000012e+00, -3.27831083e-09, 3.91493942e+00]) result=0.7950279863016516\n", - "params=array([ 1.00000013e+00, -3.27831083e-09, 3.91493942e+00]) result=0.7950279863016517\n", - "params=array([1.00000012e+00, 1.16228504e-08, 3.91493942e+00]) result=0.7950279871294111\n", - "params=array([ 1.00000012e+00, -3.27831083e-09, 3.91493943e+00]) result=0.7950279863012738\n", - "params=array([ 1.00000012e+00, -3.27641970e-09, 3.91493942e+00]) result=0.7950279863014876\n", - "params=array([ 1.00000013e+00, -3.27641970e-09, 3.91493942e+00]) result=0.7950279863014876\n", - "params=array([1.00000012e+00, 1.16247415e-08, 3.91493942e+00]) result=0.7950279871296223\n", - "params=array([ 1.00000012e+00, -3.27641970e-09, 3.91493943e+00]) result=0.7950279863011099\n", - "params=array([ 1.00000012e+00, -3.27626908e-09, 3.91493942e+00]) result=0.7950279863014748\n", - "params=array([ 1.00000013e+00, -3.27626908e-09, 3.91493942e+00]) result=0.7950279863014745\n", - "params=array([1.00000012e+00, 1.16248921e-08, 3.91493942e+00]) result=0.7950279871296392\n", - "params=array([ 1.00000012e+00, -3.27626908e-09, 3.91493943e+00]) result=0.795027986301097\n", - "params=array([ 1.00000012e+00, -3.27625698e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625698e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249042e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625698e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27626497e-09, 3.91493942e+00]) result=0.7950279863014744\n", - "params=array([ 1.00000013e+00, -3.27626497e-09, 3.91493942e+00]) result=0.7950279863014744\n", - "params=array([1.00000012e+00, 1.16248962e-08, 3.91493942e+00]) result=0.7950279871296396\n", - "params=array([ 1.00000012e+00, -3.27626497e-09, 3.91493943e+00]) result=0.7950279863010965\n", - "params=array([ 1.00000012e+00, -3.27625757e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625757e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249036e-08, 3.91493942e+00]) result=0.7950279871296403\n", - "params=array([ 1.00000012e+00, -3.27625757e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625701e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625701e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249042e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625701e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625703e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625703e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249042e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625703e-09, 3.91493943e+00]) result=0.7950279863010957\n", - "params=array([ 1.00000012e+00, -3.27625730e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625730e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296403\n", - "params=array([ 1.00000012e+00, -3.27625730e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625704e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625704e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249042e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625704e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625721e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625721e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249040e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625721e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625722e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625722e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249040e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625722e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625722e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625722e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249040e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625722e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625728e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([ 1.00000013e+00, -3.27625728e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625728e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625726e-09, 3.91493942e+00]) result=0.7950279863014736\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625726e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014738\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296404\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([ 1.00000013e+00, -3.27625727e-09, 3.91493942e+00]) result=0.7950279863014735\n", - "params=array([1.00000012e+00, 1.16249039e-08, 3.91493942e+00]) result=0.7950279871296405\n", - "params=array([ 1.00000012e+00, -3.27625727e-09, 3.91493943e+00]) result=0.7950279863010958\n", - "params=array([ 1.00000012e+00, -2.53458755e-07, 3.91506263e+00]) result=0.7950280089897778\n", - "params=array([ 1.00000012e+00, -5.69104409e-08, 3.91496583e+00]) result=0.7950279909978448\n", - "params=array([ 1.00000012e+00, -1.50051685e-08, 3.91494520e+00]) result=0.7950279873206803\n", - "params=array([ 1.00000012e+00, -5.85220054e-09, 3.91494069e+00]) result=0.7950279865249404\n", - "params=array([ 1.00000012e+00, -3.84252568e-09, 3.91493970e+00]) result=0.7950279863505801\n", - "params=array([ 1.00000012e+00, -3.40076481e-09, 3.91493948e+00]) result=0.7950279863122699\n", - "params=array([ 1.00000012e+00, -3.30363390e-09, 3.91493943e+00]) result=0.7950279863038474\n", - "params=array([ 1.00000012e+00, -3.28227628e-09, 3.91493942e+00]) result=0.7950279863019957\n", - "params=array([ 1.00000012e+00, -3.27757972e-09, 3.91493942e+00]) result=0.7950279863015884\n", - "params=array([ 1.00000012e+00, -3.27654699e-09, 3.91493942e+00]) result=0.7950279863014988\n", - "params=array([ 1.00000012e+00, -3.27631997e-09, 3.91493942e+00]) result=0.795027986301479\n", - "params=array([ 1.00000012e+00, -3.27627033e-09, 3.91493942e+00]) result=0.7950279863014749\n", "res= message: Desired error not necessarily achieved due to precision loss.\n", " success: False\n", " status: 2\n", - " fun: 0.7950279863014736\n", - " x: [ 1.000e+00 -3.276e-09 3.915e+00]\n", - " nit: 19\n", - " jac: [-7.451e-09 5.558e-02 -2.536e-05]\n", - " hess_inv: [[ 1.000e+00 -2.894e-08 -1.959e-04]\n", - " [-2.894e-08 4.363e-06 -3.025e-04]\n", - " [-1.959e-04 -3.025e-04 4.195e+00]]\n", - " nfev: 400\n", - " njev: 97\n" + " fun: 1.9265866769123492\n", + " x: [ 1.000e+00 -1.853e-09 1.923e+00]\n", + " nit: 3\n", + " jac: [-2.980e-08 2.121e+00 -1.862e+00]\n", + " hess_inv: [[ 1.000e+00 4.744e-08 1.094e-07]\n", + " [ 4.744e-08 4.993e-02 1.442e-02]\n", + " [ 1.094e-07 1.442e-02 5.616e-02]]\n", + " nfev: 388\n", + " njev: 94\n" ] }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -556,9 +156,9 @@ "def loss(params):\n", " sum = 0\n", " for i in range(len(mos_scores)):\n", - " sum += np.sum((mos_scores[i] - predict(z_scores[i], params)) ** 2) / mos_scores[i].shape[0]\n", + " err = np.sum((mos_scores[i] - predict(z_scores[i], params)) ** 2) / mos_scores[i].shape[0]\n", + " sum += err * err\n", " result = sum / len(mos_scores)\n", - " print(f'{params=} {result=}')\n", " return result\n", "\n", "\n", @@ -572,39 +172,6 @@ "plt.legend()\n", "plt.show()" ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": 19, - "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], - "source": [ - "plt.scatter(z_scores, mos_scores)\n", - "x = np.linspace(0, z_extremes[0], 1000)\n", - "plt.plot(x, predict(x, [4.568494445942755, -0.9999059785585912, 0.0006734929810393679]), \"r\", label=\"Predicted\")\n", - "plt.xlabel(\"Zimtohrli\")\n", - "plt.ylabel(\"MOS\")\n", - "plt.legend()\n", - "plt.show()" - ] } ], "metadata": {