entrenar 0.7.9

Training & Optimization library with autograd, LoRA, quantization, and model merging
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149

//! Metric and resource snapshot types for dashboard display.

use serde::{Deserialize, Serialize};

use super::Trend;
use crate::storage::MetricPoint;

/// A snapshot of metric values for dashboard display.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct MetricSnapshot {
    /// Metric key (e.g., "loss", "accuracy")
    pub key: String,
    /// Time-value pairs: (timestamp_ms, value)
    pub values: Vec<(u64, f64)>,
    /// Current trend direction
    pub trend: Trend,
}

impl MetricSnapshot {
    /// Create a new metric snapshot.
    pub fn new(key: impl Into<String>, values: Vec<(u64, f64)>) -> Self {
        let trend = Trend::from_values(&values.iter().map(|(_, v)| *v).collect::<Vec<_>>());
        Self { key: key.into(), values, trend }
    }

    /// Create from metric points.
    pub fn from_points(key: impl Into<String>, points: &[MetricPoint]) -> Self {
        let values: Vec<(u64, f64)> = points
            .iter()
            .map(|p| {
                let ts = p.timestamp.timestamp_millis() as u64;
                (ts, p.value)
            })
            .collect();
        Self::new(key, values)
    }

    /// Get the latest value.
    pub fn latest(&self) -> Option<f64> {
        self.values.last().map(|(_, v)| *v)
    }

    /// Get the minimum value.
    pub fn min(&self) -> Option<f64> {
        self.values.iter().map(|(_, v)| *v).reduce(f64::min)
    }

    /// Get the maximum value.
    pub fn max(&self) -> Option<f64> {
        self.values.iter().map(|(_, v)| *v).reduce(f64::max)
    }

    /// Get the mean value.
    pub fn mean(&self) -> Option<f64> {
        if self.values.is_empty() {
            return None;
        }
        Some(self.values.iter().map(|(_, v)| *v).sum::<f64>() / self.values.len() as f64)
    }

    /// Check if metric is empty.
    pub fn is_empty(&self) -> bool {
        self.values.is_empty()
    }

    /// Get number of data points.
    pub fn len(&self) -> usize {
        self.values.len()
    }
}

/// Resource usage snapshot for dashboard display.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct ResourceSnapshot {
    /// GPU utilization (0.0 to 1.0)
    pub gpu_util: f64,
    /// CPU utilization (0.0 to 1.0)
    pub cpu_util: f64,
    /// Memory used in bytes
    pub memory_used: u64,
    /// Total memory in bytes
    pub memory_total: u64,
    /// GPU memory used in bytes (optional)
    pub gpu_memory_used: Option<u64>,
    /// Total GPU memory in bytes (optional)
    pub gpu_memory_total: Option<u64>,
}

impl Default for ResourceSnapshot {
    fn default() -> Self {
        Self::new()
    }
}

impl ResourceSnapshot {
    /// Create a new resource snapshot with zero values.
    pub fn new() -> Self {
        Self {
            gpu_util: 0.0,
            cpu_util: 0.0,
            memory_used: 0,
            memory_total: 0,
            gpu_memory_used: None,
            gpu_memory_total: None,
        }
    }

    /// Set GPU utilization.
    pub fn with_gpu_util(mut self, util: f64) -> Self {
        self.gpu_util = util.clamp(0.0, 1.0);
        self
    }

    /// Set CPU utilization.
    pub fn with_cpu_util(mut self, util: f64) -> Self {
        self.cpu_util = util.clamp(0.0, 1.0);
        self
    }

    /// Set memory usage.
    pub fn with_memory(mut self, used: u64, total: u64) -> Self {
        self.memory_used = used;
        self.memory_total = total;
        self
    }

    /// Set GPU memory usage.
    pub fn with_gpu_memory(mut self, used: u64, total: u64) -> Self {
        self.gpu_memory_used = Some(used);
        self.gpu_memory_total = Some(total);
        self
    }

    /// Get memory utilization as a fraction.
    pub fn memory_util(&self) -> f64 {
        if self.memory_total == 0 {
            return 0.0;
        }
        self.memory_used as f64 / self.memory_total as f64
    }

    /// Get GPU memory utilization as a fraction.
    pub fn gpu_memory_util(&self) -> Option<f64> {
        match (self.gpu_memory_used, self.gpu_memory_total) {
            (Some(used), Some(total)) if total > 0 => Some(used as f64 / total as f64),
            _ => None,
        }
    }
}