/* Want to test simple MatMul example:
    [[0,0],    *   [[0,0],    =   [[0,0],
     [0,0]]         [0,0]]         [0,0]]
  */

  // Build an abstract input tensor.
  int64_t dims[] = {2, 2};  // Matrices will be 2 x 2
  int num_dims = sizeof(dims) / sizeof(dims[0]);

  float vals[] = {0.0f, 0.0f, 0.0f, 0.0f};
  TFE_Context* eager_ctx = TF_ExecutionContextGetTFEContext(ctx, status.get());
  TFE_TensorHandle* t =

  TFE_DeleteContextOptions(opts);

  TFE_ContextSetServerDef(ctx, 0, serialized.data(), serialized.size(), status);
  EXPECT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

  // Use large matrices so that RPCs don't return before we get a chance
  // to call TFE_DeleteContext.
  TFE_TensorHandle* h0_task0 = TestMatrixTensorHandle100x100(ctx);
  TFE_TensorHandle* h1_task0 = TestMatrixTensorHandle100x100(ctx);

* L'apprentissage automatique (ou **Machine Learning**) : cela nécessite de nombreuses multiplications de matrices et vecteurs. Imaginez une énorme feuille de calcul remplie de nombres que vous multiplierez entre eux tous au même moment.

* **Machine Learning**: normalmente requiere muchas multiplicaciones de "matrices" y "vectores". Imagina en una enorme hoja de cálculo con números y tener que multiplicarlos todos al mismo tiempo.

type minioCollector struct {
	desc *prometheus.Desc
}

// Describe sends the super-set of all possible descriptors of metrics
func (c *minioCollector) Describe(ch chan<- *prometheus.Desc) {
	ch <- c.desc
}

// Collect is called by the Prometheus registry when collecting metrics.
func (c *minioCollector) Collect(ch chan<- prometheus.Metric) {
	// Expose MinIO's version information

	name := rm.Name
	metrics := make([]MetricV2, 0, 3)
	metrics = append(metrics, MetricV2{
		Description:    getResourceMetricDescription(subSys, name, help),
		Value:          rm.Current,
		VariableLabels: cloneMSS(rm.Labels),
	})

	if requireAvgMax {
		avgName := MetricName(fmt.Sprintf("%s_avg", name))
		avgHelp := fmt.Sprintf("%s (avg)", help)
		metrics = append(metrics, MetricV2{

)

// Collector paths.
//
// These are paths under the top-level /minio/metrics/v3 metrics endpoint. Each
// of these paths returns a set of V3 metrics.
//
// Per-bucket metrics endpoints always start with /bucket and the bucket name is
// appended to the path. e.g. if the collector path is /bucket/api, the endpoint
// for the bucket "mybucket" would be /minio/metrics/v3/bucket/api/mybucket
const (

		}
		return nil
	}
}

// MetricsGroup - represents a group of metrics. It includes a `MetricsLoaderFn`
// function that provides a way to load the metrics from the system. The metrics
// are cached and refreshed after a given timeout.
//
// For metrics with a `bucket` dimension, a list of buckets argument is required
// to collect the metrics.
//
// It implements the prometheus.Collector interface for metric groups without a

type InQueueStats struct {
	nowBytes   int64 `json:"-"`
	nowCount   int64 `json:"-"`
	histCounts metrics.Histogram
	histBytes  metrics.Histogram
}

func newInQueueStats(r metrics.Registry, lbl string) InQueueStats {
	histCounts := metrics.NewHistogram(metrics.NewUniformSample(100))
	histBytes := metrics.NewHistogram(metrics.NewUniformSample(100))

	r.Register("replication.queue.counts."+lbl, histCounts)

				return
			}
		case "Metrics":
			var zb0002 uint32
			zb0002, err = dc.ReadMapHeader()
			if err != nil {
				err = msgp.WrapError(err, "Metrics")
				return
			}
			if z.Metrics == nil {
				z.Metrics = make(map[string]SRMetric, zb0002)
			} else if len(z.Metrics) > 0 {
				for key := range z.Metrics {
					delete(z.Metrics, key)
				}
			}
			for zb0002 > 0 {