Anima Batch Check Status

Once AnimaBatchStartTrain fires, training runs in a subprocess and you're mostly blind. AnimaBatchCheckStatus is the windshield wiper: give it the job_id the start node handed you and it reads back the job's status and the tail of its log, so you can see which folder is training, whether anything failed, and where. It's a small node, but it's the difference between "is it still going?" and "what actually happened?"

How it works

The training runner keeps a job record in a JSON file on disk - ~/.cache\comfyui_anima_batch_lora_trainer\state.json on Windows - updated with timestamps, per-task counters, a current_task field, and a rolling log (capped at 2000 lines). format_job_status in train_runner.py just reads that file and pretty-prints it. Because it's file-based rather than a live handle, the status survives even if you restart ComfyUI mid-run, and you can check on a job long after the queue has moved on.

The status line is one of queued, running, completed, failed, or partial_failed (some tasks passed, some didn't). The output header shows progress as completed/total tasks, the current folder name, and then the requested number of log lines below it.

The inputs

Only two, both trivial:

• job_id - paste in the id from AnimaBatchStartTrain's job_id output. Leave it empty and you just get ERROR: job_id is empty.
• tail_lines - how many log lines to show, default 80, up to 1000. For a long batch where you only care about the latest folder, keep it low; crank it when something failed and you want context.

The single output is status_log, a plain text string. It's marked as an output node, so it usually just sits at the end of the workflow as your readout - no downstream wiring needed.

Installing it

It's part of the AnimaForge Windows pack, so the install is the pack install: ComfyUI Manager → "ComfyUI-AnimaForge-Windows", or

cd C:\ComfyUI\custom_nodes
git clone https://github.com/AI-KSK/ComfyUI-AnimaForge-Windows.git
cd ComfyUI-AnimaForge-Windows
powershell -ExecutionPolicy Bypass -File .\scripts\install_windows.ps1

Restart ComfyUI after installing. (The full -SetupBackend training backend is only needed once you actually train - this node itself is just a file reader.)

Common issues

The main way people get burned is feeding it a stale or wrong job_id - you get ERROR: job not found and, because each run of AnimaBatchStartTrain mints a new id, an old id stays valid but points at an old job. If you re-ran the start node, grab the new id. There's also a genuine gotcha in how ComfyUI executes: since the start node blocks the queue until training finishes, this status node in the same workflow runs after the batch completes - so during a long run it won't update live in the graph. Check the state file or a second workflow when you want mid-run visibility. And remember that a partial_failed status isn't a crash: the batch kept going, and the failed folder's own .train.log in its output directory is where the real error text lives.