jtanios

Resume training from a saved checkpoint

sym, arg_params, aux_params = mx.model.load_checkpoint(model_prefix, epoch)
model = mx.model.FeedForward(symbol=sym, arg_params=arg_params, aux_params=aux_params, ctx=devices, ...)
model.fit(...)

Fine-tuning a model (transfer learning)

Functions in fine_tune.py

Quick examples

feat: new feature
fix(scope): bug in scope
feat!: breaking change / feat(scope)!: change API contract
chore(deps): update dependencies

Commit types

feat: A new product feature
fix: A bug fix
test: Adding tests or correcting existing tests

	-- Shows process IDs using MyDatabase
	SELECT SPId FROM master..SysProcesses
	WHERE DBId = DB_ID('MyDatabase') AND SPId <> @@SPID

	-- Run this for each process ID
	KILL <process_id>

	-- Take offline
	ALTER DATABASE [MyDatabase] SET OFFLINE WITH ROLLBACK IMMEDIATE