File size: 5,142 Bytes
9f818c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: OpenMDW-1.1

from __future__ import annotations

from dataclasses import dataclass
from typing import Tuple

import torch
import torch.distributed as dist
import torch.utils.data
import wandb

from cosmos_framework.model._base import ImaginaireModel
from cosmos_framework.utils import distributed, log
from cosmos_framework.utils.callback import Callback
from cosmos_framework.utils.easy_io import easy_io


@dataclass
class _LossRecord:
    loss: float = 0
    iter_count: int = 0

    def reset(self) -> None:
        self.loss = 0
        self.iter_count = 0

    def get_stat(self) -> Tuple[float, float]:
        if self.iter_count > 0:
            avg_loss_tensor = self.loss / self.iter_count
            # Create a mask (1 if valid, 0 if NaN or Inf)
            valid_mask = torch.tensor([torch.isfinite(avg_loss_tensor).float()], device="cuda")

            # Replace NaN/Inf with 0 to avoid affecting sum
            avg_loss_tensor = torch.where(
                torch.isfinite(avg_loss_tensor), avg_loss_tensor, torch.tensor([0.0], device="cuda")
            )

            # Reduce across all ranks
            dist.all_reduce(avg_loss_tensor, op=dist.ReduceOp.SUM)  # Sum of valid losses
            dist.all_reduce(valid_mask, op=dist.ReduceOp.SUM)  # Count of valid losses

            # Compute final average, avoiding division by zero
            if valid_mask.item() > 0:
                final_avg_loss = (avg_loss_tensor / valid_mask).item()
            else:
                final_avg_loss = 0.0  # Default to zero if all values were invalid

            avg_loss = final_avg_loss
        else:
            avg_loss = 0
        self.reset()
        return avg_loss


class WandbCallback(Callback):
    def __init__(
        self,
        save_s3: bool = False,
    ) -> None:
        super().__init__()
        self.final_loss_log = _LossRecord()
        self.final_loss_log_per_dataset = {}

        self.save_s3 = save_s3
        self.wandb_extra_tag = ""
        self.name = "wandb_loss_val_log"
        self.unstable_count = torch.zeros(1, device="cuda")
        self.url_key_list = []

    def on_validation_step_end(
        self,
        model: ImaginaireModel,
        data_batch: dict[str, torch.Tensor],
        output_batch: dict[str, torch.Tensor],
        loss: torch.Tensor,
        iteration: int = 0,
    ) -> None:
        if torch.isnan(loss) or torch.isinf(loss):
            log.critical(
                f"Unstable val loss {loss} at iteration {iteration}",
                rank0_only=False,
            )
            self.unstable_count += 1

        dataset_name = data_batch.get("dataset_name", "default")

        # Handle case where dataset_name gets batched into a list
        if isinstance(dataset_name, list):

            assert len(dataset_name) == 1, "dataset_name should be a list of 1"
            dataset_name = dataset_name[0]

        if dataset_name not in self.final_loss_log_per_dataset:
            self.final_loss_log_per_dataset[dataset_name] = _LossRecord()

        self.final_loss_log_per_dataset[dataset_name].loss += loss.detach().float()
        self.final_loss_log_per_dataset[dataset_name].iter_count += 1
        self.final_loss_log.loss += loss.detach().float()
        self.final_loss_log.iter_count += 1

        self.url_key_list.append(f"{data_batch.get('__url__', [''])[0]}, {data_batch.get('__key__', [''])[0]}")

    def on_validation_end(self, model: ImaginaireModel, iteration: int = 0) -> None:
        avg_final_loss = self.final_loss_log.get_stat()

        log.info(f"avg_final_loss: {avg_final_loss}")
        dist.all_reduce(self.unstable_count, op=dist.ReduceOp.SUM)
        # gather url and key list from all ranks
        url_key_list = [None] * dist.get_world_size()
        dist.all_gather_object(url_key_list, self.url_key_list)
        url_key_list = [item for sublist in url_key_list for item in sublist]

        unique_url_key_list = list(set(url_key_list))
        if distributed.is_rank0():
            info = {}
            log.info(
                f"[val] number of unique url and key: {len(unique_url_key_list)} / {len(url_key_list)}; avg_final_loss: {avg_final_loss}"
            )
            info.update(
                {
                    f"val{self.wandb_extra_tag}/loss": avg_final_loss,
                    f"val{self.wandb_extra_tag}/unstable_count": self.unstable_count.item(),
                    "iteration": iteration,
                    f"val{self.wandb_extra_tag}/num_unique_url_key": len(unique_url_key_list),
                    f"val{self.wandb_extra_tag}/total_url_key": len(url_key_list),
                }
            )
            if self.save_s3:
                easy_io.dump(
                    info,
                    f"s3://rundir/{self.name}/Val_Iter{iteration:09d}.json",
                )

            if wandb.run is not None:
                wandb.log(info, step=iteration)

        # reset unstable count
        self.unstable_count.zero_()
        self.url_key_list = []