# Copyright 2023 Big Vision Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """Training loop for CapPa (https://arxiv.org/abs/2306.07915).""" # pylint: disable=consider-using-from-import import functools import importlib import multiprocessing.pool import os from absl import app from absl import flags from absl import logging import big_vision.datasets.core as ds_core import big_vision.evaluators.common as eval_common import big_vision.input_pipeline as input_pipeline import big_vision.optax as bv_optax import big_vision.sharding as bv_sharding import big_vision.trainers.proj.cappa.predict_fns as predict_fns import big_vision.utils as u from clu import parameter_overview import flax import flax.linen as nn import jax from jax.experimental import mesh_utils from jax.experimental import multihost_utils from jax.experimental.array_serialization import serialization as array_serial import jax.numpy as jnp from ml_collections import config_flags import numpy as np import optax import tensorflow as tf from tensorflow.io import gfile # pylint: disable=logging-fstring-interpolation config_flags.DEFINE_config_file( "config", None, "Training configuration.", lock_config=True) flags.DEFINE_string("workdir", default=None, help="Work unit directory.") flags.DEFINE_boolean("cleanup", default=False, help="Delete workdir (only) after successful completion.") # Adds jax flags to the program. jax.config.parse_flags_with_absl() # Transfer guard will fail the program whenever that data between a host and # a device is transferred implicitly. This often catches subtle bugs that # cause slowdowns and memory fragmentation. Explicit transfers are done # with jax.device_put and jax.device_get. jax.config.update("jax_transfer_guard", "disallow") # Fixes design flaw in jax.random that may cause unnecessary d2d comms. jax.config.update("jax_threefry_partitionable", True) def main(argv): del argv try: jax.distributed.initialize() except ValueError as e: logging.warning('Could not initialize distributed environment: %s', e) # Make sure TF does not touch GPUs. tf.config.set_visible_devices([], "GPU") # Consistent device order is important to ensure correctness of various train # loop components, such as input pipeline, update step, evaluators. We use # jax utils to infer device order that will be used throughout the program. config = flags.FLAGS.config num_sharded_replicas = config.get("num_sharded_replicas", 1) assert jax.device_count() % num_sharded_replicas == 0, ( num_sharded_replicas, jax.device_count()) devices = mesh_utils.create_device_mesh( (num_sharded_replicas, jax.device_count() // num_sharded_replicas)) devices_flat = devices.reshape(-1) ################################################################################ # # # Set up logging # # # ################################################################################ # Set up work directory and print welcome message. config = flags.FLAGS.config workdir = flags.FLAGS.workdir logging.info( f"\u001b[33mHello from process {jax.process_index()} holding " f"{jax.local_device_count()}/{jax.device_count()} devices and " f"writing to workdir {workdir}.\u001b[0m") save_ckpt_path = None if workdir: # Always create if requested, even if we may not write into it. gfile.makedirs(workdir) save_ckpt_path = os.path.join(workdir, "checkpoint.bv") # The pool is used to perform misc operations such as logging in async way. pool = multiprocessing.pool.ThreadPool() # Here we register preprocessing ops from modules listed on `pp_modules`. for m in config.get("pp_modules", ["ops_general", "ops_image", "ops_text"]): importlib.import_module(f"big_vision.pp.{m}") # Setup up logging and experiment manager. xid, wid = -1, -1 fillin = lambda s: s def info(s, *a): logging.info("\u001b[33mNOTE\u001b[0m: " + s, *a) def write_note(note): if jax.process_index() == 0: info("%s", note) mw = u.BigVisionMetricWriter(xid, wid, workdir, config) ################################################################################ # # # Input Pipeline # # # ################################################################################ write_note("Initializing train dataset...") batch_size = config.input.batch_size if batch_size % jax.device_count() != 0: raise ValueError(f"Batch size ({batch_size}) must " f"be divisible by device number ({jax.device_count()})") info("Global batch size %d on %d hosts results in %d local batch size. With " "%d dev per host (%d dev total), that's a %d per-device batch size.", batch_size, jax.process_count(), batch_size // jax.process_count(), jax.local_device_count(), jax.device_count(), batch_size // jax.device_count()) train_ds, ntrain_img = input_pipeline.training(config.input) total_steps = u.steps("total", config, ntrain_img, batch_size) def get_steps(name, default=ValueError, cfg=config): return u.steps(name, cfg, ntrain_img, batch_size, total_steps, default) u.chrono.inform(total_steps=total_steps, global_bs=batch_size, steps_per_epoch=ntrain_img / batch_size, measure=mw.measure, write_note=write_note) info("Running for %d steps, that means %f epochs", total_steps, total_steps * batch_size / ntrain_img) # Start input pipeline as early as possible. n_prefetch = config.get("prefetch_to_device", 1) train_iter = input_pipeline.start_global(train_ds, devices_flat, n_prefetch) # For mixed data, add per-dataset epoch and examples seen measurements. if isinstance(config.input.data.get("name"), str): measure_per_dataset_times = lambda step: None # No-op else: nexamples = { name: ds_core.get(**config.input[name].data).total_examples for name in config.input.data } def measure_per_dataset_times(step): total = sum(config.input.data.values()) for name, w in config.input.data.items(): w = w / total mw.measure(f"examples_seen_{name}", u.chrono.accum_examples_seen * w) mw.measure(f"epoch_{name}", step * batch_size * w / nexamples[name]) ################################################################################ # # # Create Model & Optimizer # # # ################################################################################ write_note(f"Initializing {config.model_name} model...") model_mod = importlib.import_module(f"big_vision.models.{config.model_name}") model = model_mod.Model(**config.get("model", {})) def init(rng): bs = batch_size // jax.device_count() img_shape = (bs,) + tuple(train_ds.element_spec["image"].shape[1:]) txt_shape = (bs,) + tuple(train_ds.element_spec["labels"].shape[1:]) dummy_img = jnp.zeros(img_shape, jnp.float32) dummy_txt = jnp.zeros(txt_shape, jnp.int64) variables = model.init(rng, dummy_img, dummy_txt) return flax.core.unfreeze(variables["params"]) # This seed makes the Jax part of things (like model init) deterministic. # However, full training still won't be deterministic, for example due to the # tf.data pipeline not being deterministic even if we would set TF seed. # See (internal link) for a fun read on what it takes. rng = jax.random.PRNGKey(u.put_cpu(config.get("seed", 0))) rng, rng_init = jax.random.split(rng) with u.chrono.log_timing("z/secs/init"): write_note("Inferring parameter shapes...") params_shape = jax.eval_shape(init, rng_init) write_note("Inferring optimizer state shapes...") tx, sched_fns = bv_optax.make(config, params_shape, sched_kw=dict( total_steps=total_steps, batch_size=batch_size, data_size=ntrain_img)) opt_shape = jax.eval_shape(tx.init, params_shape) # We jit this, such that the arrays are created on the CPU, not device[0]. sched_fns_cpu = [u.jit_cpu()(sched_fn) for sched_fn in sched_fns] num_params = sum(np.prod(p.shape) for p in jax.tree_leaves(params_shape)) mw.measure("num_params", num_params) ################################################################################ # # # Shard & Transfer # # # ################################################################################ # Using a 2D mesh where the model/optimizer parameters are replicated along # the "replica" axis and sharded along the "fsdp" axis if the sharding is set # to "fully_sharded" for the model/optimizer. write_note("Creating device mesh...") mesh = jax.sharding.Mesh(devices, ("replica", "fsdp")) repl_sharding = jax.sharding.NamedSharding(mesh, jax.sharding.PartitionSpec()) write_note("Inferring shardings...") params_sharding = bv_sharding.infer_sharding( params_shape, mesh, axis_name="fsdp", # TODO: implement scan for parameter sharding. strategy=config.get("param_sharding", "replicated"), extra_strategy_args=config.get("param_sharding_args", {})) opt_sharding = bv_sharding.infer_sharding( opt_shape, mesh, axis_name="fsdp", strategy=config.get("optim_sharding", "replicated"), extra_strategy_args=config.get("optim_sharding_args", {})) write_note("Transferring train_state to devices...") # RNG is always replicated rng_init = u.reshard(rng_init, repl_sharding) # Parameters and the optimizer are now global (distributed) jax arrays. params = jax.jit(init, out_shardings=params_sharding)(rng_init) opt = jax.jit(tx.init, out_shardings=opt_sharding)(params) rng, rng_loop = jax.random.split(rng, 2) rng_loop = u.reshard(rng_loop, repl_sharding) del rng # not used anymore, so delete it. # At this point we have everything we need to form a train state. It contains # all the parameters that are passed and updated by the main training step. train_state_sharding = { "params": params_sharding, "opt": opt_sharding, "rng": repl_sharding} train_state = { "params": params, "opt": opt, "rng": rng_loop} del params, opt, rng_loop # Delete to avoid memory leak or accidental reuse. write_note("Logging parameter overview...") parameter_overview.log_parameter_overview( train_state["params"], msg="Init params", include_stats="global", jax_logging_process=0) ################################################################################ # # # Update Step # # # ################################################################################ @functools.partial( jax.jit, donate_argnums=(0,), out_shardings=(train_state_sharding, repl_sharding)) def update_fn(train_state, batch): """Update step.""" images, labels, label_masks = ( batch["image"], batch["labels"], batch.get("label_masks")) # Get device-specific loss rng. rng = train_state["rng"] rng, rng_model = jax.random.split(rng, 2) def loss_fn(params): logits = model.apply( {"params": params}, images, labels, train=True, rngs={"dropout": rng_model}) weights = jnp.where(labels != config.get("pad_token", 0), 1.0, 0.0) if label_masks is not None: weights = weights * label_masks loss = u.weighted_softmax_xent( logits=logits, labels=labels, weights=weights, label_smoothing=config.get("label_smoothing", 0.0), reduction=True, normalize=True) return loss params, opt = train_state["params"], train_state["opt"] loss, grads = jax.value_and_grad(loss_fn)(params) updates, opt = tx.update(grads, opt, params) params = optax.apply_updates(params, updates) measurements = {"training_loss": loss} gs = jax.tree_leaves(bv_optax.replace_frozen(config.schedule, grads, 0.)) measurements["l2_grads"] = jnp.sqrt(sum([jnp.sum(g * g) for g in gs])) ps = jax.tree_leaves(params) measurements["l2_params"] = jnp.sqrt(sum([jnp.sum(p * p) for p in ps])) us = jax.tree_leaves(updates) measurements["l2_updates"] = jnp.sqrt(sum([jnp.sum(u * u) for u in us])) return {"params": params, "opt": opt, "rng": rng}, measurements ################################################################################ # # # Load Checkpoint # # # ################################################################################ # Decide how to initialize training. The order is important. # 1. Always resumes from the existing checkpoint, e.g. resumes a finetune job. # 2. Resume from a previous checkpoint, e.g. start a cooldown training job. # 3. Initialize model from something, e,g, start a fine-tuning job. # 4. Train from scratch. resume_ckpt_path = None if save_ckpt_path and gfile.exists(f"{save_ckpt_path}-LAST"): resume_ckpt_path = save_ckpt_path elif config.get("resume"): resume_ckpt_path = fillin(config.resume) ckpt_mngr = None if save_ckpt_path or resume_ckpt_path: ckpt_mngr = array_serial.GlobalAsyncCheckpointManager() if resume_ckpt_path: write_note(f"Resuming training from checkpoint {resume_ckpt_path}...") jax.tree_map(lambda x: x.delete(), train_state) del train_state shardings = { **train_state_sharding, "chrono": jax.tree_map(lambda _: repl_sharding, u.chrono.save()), } loaded = u.load_checkpoint_ts( resume_ckpt_path, tree=shardings, shardings=shardings) train_state = {key: loaded[key] for key in train_state_sharding.keys()} u.chrono.load(jax.device_get(loaded["chrono"])) del loaded elif config.get("model_init"): write_note(f"Initialize model from {config.model_init}...") train_state["params"] = model_mod.load( train_state["params"], config.model_init, config.get("model"), **config.get("model_load", {})) # load has the freedom to return params not correctly sharded. Think of for # example ViT resampling position embedings on CPU as numpy arrays. train_state["params"] = u.reshard( train_state["params"], train_state_sharding["params"]) parameter_overview.log_parameter_overview( train_state["params"], msg="restored params", include_stats="global", jax_logging_process=0) ################################################################################ # # # Setup Evals # # # ################################################################################ # Only initialize evaluators when they are first needed. @functools.lru_cache(maxsize=None) def evaluators(): return eval_common.from_config( config, predict_fns.get_predict_fns(model), lambda s: write_note(f"Init evaluator: {s}…\n{u.chrono.note}"), lambda key, cfg: get_steps(key, default=None, cfg=cfg), devices_flat, ) # At this point we need to know the current step to see whether to run evals. write_note("Inferring the first step number...") first_step_device = bv_optax.get_count(train_state["opt"], jittable=True) first_step = int(jax.device_get(first_step_device)) u.chrono.inform(first_step=first_step) # Note that training can be pre-empted during the final evaluation (i.e. # just after the final checkpoint has been written to disc), in which case we # want to run the evals. if first_step in (total_steps, 0): write_note("Running initial or final evals...") mw.step_start(first_step) for (name, evaluator, _, prefix) in evaluators(): if config.evals[name].get("skip_first") and first_step != total_steps: continue write_note(f"{name} evaluation...\n{u.chrono.note}") with u.chrono.log_timing(f"z/secs/eval/{name}"): with mesh, nn.logical_axis_rules([("act_batch", ("replica", "fsdp"))]): # pytype: disable=wrong-arg-types for key, value in evaluator.run(train_state): mw.measure(f"{prefix}{key}", value) ################################################################################ # # # Train Loop # # # ################################################################################ prof = None # Keeps track of start/stop of profiler state. write_note("Starting training loop, compiling the first step...") for step, batch in zip(range(first_step + 1, total_steps + 1), train_iter): mw.step_start(step) with jax.profiler.StepTraceAnnotation("train_step", step_num=step): with u.chrono.log_timing("z/secs/update0", noop=step > first_step + 1): with mesh, nn.logical_axis_rules([("act_batch", ("replica", "fsdp"))]): # pytype: disable=wrong-arg-types train_state, measurements = update_fn(train_state, batch) # On the first host, let's always profile a handful of early steps. if jax.process_index() == 0: prof = u.startstop_prof(prof, step, first_step, get_steps("log_training")) # Report training progress if (u.itstime(step, get_steps("log_training"), total_steps, host=0) or u.chrono.warmup and jax.process_index() == 0): for i, sched_fn_cpu in enumerate(sched_fns_cpu): mw.measure(f"global_schedule{i if i else ''}", sched_fn_cpu(u.put_cpu(step - 1))) measurements = jax.device_get(measurements) for name, value in measurements.items(): mw.measure(name, value) u.chrono.tick(step) measure_per_dataset_times(step) if not np.isfinite(measurements["training_loss"]): raise RuntimeError(f"The loss became nan or inf somewhere within steps " f"[{step - get_steps('log_training')}, {step}]") # Checkpoint saving keep_ckpt_steps = get_steps("keep_ckpt", None) or total_steps if save_ckpt_path and ( (keep := u.itstime(step, keep_ckpt_steps, total_steps, first=False)) or u.itstime(step, get_steps("ckpt", None), total_steps, first=True) ): u.chrono.pause(wait_for=train_state) # Copy because we add extra stuff to the checkpoint. ckpt = {**train_state} # To save chrono state correctly and safely in a multihost setup, we # broadcast the state to all hosts and convert it to a global array. with jax.transfer_guard("allow"): chrono_ckpt = multihost_utils.broadcast_one_to_all(u.chrono.save()) chrono_shardings = jax.tree_map(lambda _: repl_sharding, chrono_ckpt) ckpt = ckpt | {"chrono": u.reshard(chrono_ckpt, chrono_shardings)} u.save_checkpoint_ts(ckpt_mngr, ckpt, save_ckpt_path, step, keep) u.chrono.resume() for (name, evaluator, log_steps, prefix) in evaluators(): if u.itstime(step, log_steps, total_steps, first=False, last=True): u.chrono.pause(wait_for=train_state) u.chrono.tick(step) # Record things like epoch number, core hours etc. write_note(f"{name} evaluation...\n{u.chrono.note}") with u.chrono.log_timing(f"z/secs/eval/{name}"): with mesh, nn.logical_axis_rules( [("act_batch", ("replica", "fsdp"))]): # pytype: disable=wrong-arg-types for key, value in evaluator.run(train_state): mw.measure(f"{prefix}{key}", jax.device_get(value)) u.chrono.resume() mw.step_end() # Always give a chance to stop the profiler, no matter how things ended. # TODO: can we also do this when dying of an exception like OOM? if jax.process_index() == 0 and prof is not None: u.startstop_prof(prof) # Last note needs to happen before the pool's closed =) write_note(f"Done!\n{u.chrono.note}") pool.close() pool.join() mw.close() if ckpt_mngr: ckpt_mngr.wait_until_finished() # Make sure all hosts stay up until the end of main. u.sync() u.maybe_cleanup_workdir(workdir, flags.FLAGS.cleanup, info) if __name__ == "__main__": app.run(main)