tensorlogic-train 0.1.0

Training loops, loss composition, and optimization schedules for TensorLogic
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

//! LoRA-specific error types.

use thiserror::Error;

#[derive(Error, Debug)]
pub enum LoraError {
    #[error("invalid rank: {0} (must be >= 1 and <= min(d, k))")]
    InvalidRank(usize),

    #[error("dimension mismatch: expected {expected}, got {got}")]
    DimensionMismatch { expected: String, got: String },

    #[error("merge error: {0}")]
    MergeError(String),

    #[error("frozen weights: {0}")]
    FrozenWeights(String),
}

pub type LoraResult<T> = Result<T, LoraError>;