LICENSE
README.md
pyproject.toml
requirements-dev.txt
requirements.txt
.ci/docker/requirements-flux.txt
.ci/docker/requirements-rl.txt
.ci/docker/requirements-vlm.txt
assets/version.txt
torchtitan/__init__.py
torchtitan/train.py
torchtitan/trainer.py
torchtitan/training_engine.py
torchtitan.egg-info/PKG-INFO
torchtitan.egg-info/SOURCES.txt
torchtitan.egg-info/dependency_links.txt
torchtitan.egg-info/requires.txt
torchtitan.egg-info/top_level.txt
torchtitan/_torch_compat/__init__.py
torchtitan/_torch_compat/_pipelining_edits.py
torchtitan/components/loss.py
torchtitan/components/renderer.py
torchtitan/components/tokenizer.py
torchtitan/components/validator.py
torchtitan/components/checkpointer/__init__.py
torchtitan/components/checkpointer/base.py
torchtitan/components/checkpointer/dcp.py
torchtitan/components/checkpointer/torch_checkpointing.py
torchtitan/components/checkpointer/utils.py
torchtitan/components/data/__init__.py
torchtitan/components/data/collators.py
torchtitan/components/data/dataset.py
torchtitan/components/data/loader.py
torchtitan/components/data/packing.py
torchtitan/components/data/sources.py
torchtitan/components/data/text_datasets.py
torchtitan/components/data/types.py
torchtitan/components/data/multimodal/__init__.py
torchtitan/components/data/multimodal/mm_collator.py
torchtitan/components/data/multimodal/mm_datasets.py
torchtitan/components/data/multimodal/utils/audio.py
torchtitan/components/data/multimodal/utils/image.py
torchtitan/components/data/multimodal/utils/text.py
torchtitan/components/data/multimodal/utils/video.py
torchtitan/components/optim/__init__.py
torchtitan/components/optim/ema.py
torchtitan/components/optim/lr_scheduler.py
torchtitan/components/optim/optim.py
torchtitan/components/optim/optimizer.py
torchtitan/components/optim/utils.py
torchtitan/config/__init__.py
torchtitan/config/configs.py
torchtitan/config/configurable.py
torchtitan/config/function.py
torchtitan/config/loader.py
torchtitan/config/override.py
torchtitan/config/parallelism.py
torchtitan/config/validation.py
torchtitan/config/transform/__init__.py
torchtitan/config/transform/apply.py
torchtitan/config/transform/async_tensor_parallel.py
torchtitan/config/transform/base.py
torchtitan/config/transform/batch_invariance.py
torchtitan/config/transform/context_parallel.py
torchtitan/config/transform/converter.py
torchtitan/config/transform/dist_moe.py
torchtitan/config/transform/freeze_moe_routing.py
torchtitan/config/transform/lm_head_fp32.py
torchtitan/config/transform/lora.py
torchtitan/config/transform/quantization.py
torchtitan/config/transform/relations.py
torchtitan/config/transform/token_dispatcher.py
torchtitan/distributed/__init__.py
torchtitan/distributed/activation_remat.py
torchtitan/distributed/batch_invariant.py
torchtitan/distributed/context_parallel.py
torchtitan/distributed/cuda_graph.py
torchtitan/distributed/fsdp.py
torchtitan/distributed/local_compile.py
torchtitan/distributed/offloading.py
torchtitan/distributed/parallelism_context.py
torchtitan/distributed/pipeline_parallel.py
torchtitan/distributed/spmd_types.py
torchtitan/distributed/utils.py
torchtitan/distributed/deepep/deepep.py
torchtitan/distributed/deepep/hybridep.py
torchtitan/distributed/flex_shard/__init__.py
torchtitan/distributed/flex_shard/_optimizer_reshard_runtime.py
torchtitan/distributed/flex_shard/_optimizer_reshard_schedule.py
torchtitan/distributed/flex_shard/dist_muon.py
torchtitan/distributed/flex_shard/optimizer_reshard.py
torchtitan/experiments/__init__.py
torchtitan/experiments/graph_trainer/__init__.py
torchtitan/experiments/graph_trainer/autoparallel_api.py
torchtitan/experiments/graph_trainer/chunked_loss.py
torchtitan/experiments/graph_trainer/common_utils.py
torchtitan/experiments/graph_trainer/compile.py
torchtitan/experiments/graph_trainer/configs.py
torchtitan/experiments/graph_trainer/cpu_offload.py
torchtitan/experiments/graph_trainer/cuda_graph.py
torchtitan/experiments/graph_trainer/custom_codegen.py
torchtitan/experiments/graph_trainer/debug_utils.py
torchtitan/experiments/graph_trainer/decompositions.py
torchtitan/experiments/graph_trainer/dynamic_shapes.py
torchtitan/experiments/graph_trainer/ep_eager_chunk.py
torchtitan/experiments/graph_trainer/ep_overlap_pass.py
torchtitan/experiments/graph_trainer/ep_pass_utils.py
torchtitan/experiments/graph_trainer/ep_process_group_pass.py
torchtitan/experiments/graph_trainer/fsdp_passes.py
torchtitan/experiments/graph_trainer/fsdp_patterns.py
torchtitan/experiments/graph_trainer/grad_accumulation.py
torchtitan/experiments/graph_trainer/graph_builder.py
torchtitan/experiments/graph_trainer/graph_builder_utils.py
torchtitan/experiments/graph_trainer/inductor_passes.py
torchtitan/experiments/graph_trainer/log_activation_memory_policy.py
torchtitan/experiments/graph_trainer/make_fx_tracer.py
torchtitan/experiments/graph_trainer/memory_policy.py
torchtitan/experiments/graph_trainer/model.py
torchtitan/experiments/graph_trainer/mutation_utils.py
torchtitan/experiments/graph_trainer/paged_stash_memory_policy.py
torchtitan/experiments/graph_trainer/passes.py
torchtitan/experiments/graph_trainer/performance_passes.py
torchtitan/experiments/graph_trainer/precompile.py
torchtitan/experiments/graph_trainer/precompile_main.py
torchtitan/experiments/graph_trainer/registry.py
torchtitan/experiments/graph_trainer/remove_noop_passes.py
torchtitan/experiments/graph_trainer/selective_activation_remat.py
torchtitan/experiments/graph_trainer/simple_fsdp.py
torchtitan/experiments/graph_trainer/spmd_gradient_accumulation_graph_builder.py
torchtitan/experiments/graph_trainer/spmd_graph_builder.py
torchtitan/experiments/graph_trainer/storage.py
torchtitan/experiments/graph_trainer/subgraph_regions.py
torchtitan/experiments/graph_trainer/trainer.py
torchtitan/experiments/graph_trainer/wgrad_accumulation.py
torchtitan/experiments/graph_trainer/deepseek_v3/__init__.py
torchtitan/experiments/graph_trainer/deepseek_v3/flavors.py
torchtitan/experiments/graph_trainer/deepseek_v3/model.py
torchtitan/experiments/graph_trainer/deepseek_v3/parallelize_autoparallel.py
torchtitan/experiments/graph_trainer/graph_pp/__init__.py
torchtitan/experiments/graph_trainer/graph_pp/graph_multiplex.py
torchtitan/experiments/graph_trainer/graph_pp/partition.py
torchtitan/experiments/graph_trainer/graph_pp/pipeline.py
torchtitan/experiments/graph_trainer/graph_pp/pp_graph_builder.py
torchtitan/experiments/graph_trainer/graph_pp/runner.py
torchtitan/experiments/graph_trainer/graph_pp/split_di_dw.py
torchtitan/experiments/graph_trainer/graph_pp/split_fsdp_collectives.py
torchtitan/experiments/graph_trainer/graph_pp/stage.py
torchtitan/experiments/graph_trainer/graph_pp/stage_builder.py
torchtitan/experiments/graph_trainer/graph_pp/utils.py
torchtitan/experiments/graph_trainer/llama3/__init__.py
torchtitan/experiments/graph_trainer/llama3/flavors.py
torchtitan/experiments/graph_trainer/llama3/model.py
torchtitan/experiments/graph_trainer/llama3/parallelize_autoparallel.py
torchtitan/experiments/graph_trainer/muse_glimmer/__init__.py
torchtitan/experiments/graph_trainer/muse_glimmer/flavors.py
torchtitan/experiments/graph_trainer/muse_glimmer/model.py
torchtitan/experiments/graph_trainer/qwen3/__init__.py
torchtitan/experiments/graph_trainer/qwen3/flavors.py
torchtitan/experiments/graph_trainer/qwen3/model.py
torchtitan/experiments/graph_trainer/tests/__init__.py
torchtitan/experiments/graph_trainer/tests/_trainer_test_utils.py
torchtitan/experiments/graph_trainer/tests/integration_tests.py
torchtitan/experiments/graph_trainer/tests/run_precompile_tests.py
torchtitan/experiments/graph_trainer/tests/test_autoparallel.py
torchtitan/experiments/graph_trainer/tests/test_bitwise_deterministic.py
torchtitan/experiments/graph_trainer/tests/test_chunked_loss.py
torchtitan/experiments/graph_trainer/tests/test_cpu_offload.py
torchtitan/experiments/graph_trainer/tests/test_custom_codegen.py
torchtitan/experiments/graph_trainer/tests/test_dist_moe.py
torchtitan/experiments/graph_trainer/tests/test_grad_accumulation.py
torchtitan/experiments/graph_trainer/tests/test_graph_pp_passes.py
torchtitan/experiments/graph_trainer/tests/test_graph_pp_runner.py
torchtitan/experiments/graph_trainer/tests/test_inplace_wgrad_accum_config.py
torchtitan/experiments/graph_trainer/tests/test_numerics.py
torchtitan/experiments/graph_trainer/tests/test_paged_stash.py
torchtitan/experiments/graph_trainer/tests/test_passes.py
torchtitan/experiments/graph_trainer/tests/test_performance_passes.py
torchtitan/experiments/graph_trainer/tests/test_precompile.py
torchtitan/experiments/graph_trainer/tests/test_profiler.py
torchtitan/experiments/graph_trainer/tests/test_remat_metadata.py
torchtitan/experiments/graph_trainer/tests/test_sac_peak_memory.py
torchtitan/experiments/graph_trainer/tests/test_simple_fsdp.py
torchtitan/experiments/graph_trainer/tests/test_subgraph_regions.py
torchtitan/experiments/graph_trainer/tests/test_trace_module.py
torchtitan/experiments/graph_trainer/tests/test_wgrad_accumulation.py
torchtitan/experiments/torchft/__init__.py
torchtitan/experiments/torchft/checkpoint.py
torchtitan/experiments/torchft/manager.py
torchtitan/experiments/torchft/optimizer.py
torchtitan/experiments/torchft/trainer.py
torchtitan/experiments/torchft/config/__init__.py
torchtitan/experiments/torchft/config/job_config.py
torchtitan/experiments/torchft/diloco/__init__.py
torchtitan/experiments/torchft/diloco/utils.py
torchtitan/experiments/torchft/llama3/__init__.py
torchtitan/experiments/torchft/llama3/flavors.py
torchtitan/experiments/torchft/llama3/model.py
torchtitan/experiments/torchft/tests/__init__.py
torchtitan/experiments/torchft/tests/integration_tests.py
torchtitan/experiments/torchft/tests/test_torchft_checkpoint.py
torchtitan/experiments/torchft/tests/test_torchft_manager.py
torchtitan/experiments/torchft/tests/test_torchft_optimizer.py
torchtitan/experiments/torchft/tests/test_torchft_trainer.py
torchtitan/experiments/transformers_modeling_backend/__init__.py
torchtitan/experiments/transformers_modeling_backend/flavors.py
torchtitan/experiments/transformers_modeling_backend/hf_sharding.py
torchtitan/experiments/transformers_modeling_backend/model.py
torchtitan/experiments/transformers_modeling_backend/module_conversion.py
torchtitan/experiments/transformers_modeling_backend/moe_replacement.py
torchtitan/experiments/transformers_modeling_backend/parallelize.py
torchtitan/experiments/transformers_modeling_backend/pipeline.py
torchtitan/experiments/transformers_modeling_backend/state_dict_adapter.py
torchtitan/experiments/transformers_modeling_backend/tokenizer.py
torchtitan/experiments/transformers_modeling_backend/tests/__init__.py
torchtitan/experiments/transformers_modeling_backend/tests/cp_pp_numerical.py
torchtitan/experiments/transformers_modeling_backend/tests/integration_tests.py
torchtitan/experiments/transformers_modeling_backend/tests/test_flex_attention.py
torchtitan/experiments/transformers_modeling_backend/tests/test_flex_cp_numerical.py
torchtitan/experiments/transformers_modeling_backend/tests/test_hf_build_forward_inputs.py
torchtitan/experiments/transformers_modeling_backend/tests/test_hf_sharding.py
torchtitan/experiments/transformers_modeling_backend/tests/test_moe_parallelism.py
torchtitan/experiments/transformers_modeling_backend/tests/test_rope_init.py
torchtitan/models/__init__.py
torchtitan/models/base/__init__.py
torchtitan/models/base/model.py
torchtitan/models/base/module.py
torchtitan/models/base/sharding.py
torchtitan/models/base/state_dict_adapter.py
torchtitan/models/common/__init__.py
torchtitan/models/common/activation.py
torchtitan/models/common/attn_gym.py
torchtitan/models/common/aux_loss.py
torchtitan/models/common/config_utils.py
torchtitan/models/common/decoder.py
torchtitan/models/common/decoder_sharding.py
torchtitan/models/common/embedding.py
torchtitan/models/common/feed_forward.py
torchtitan/models/common/flops.py
torchtitan/models/common/lora.py
torchtitan/models/common/multimodal.py
torchtitan/models/common/nn_modules.py
torchtitan/models/common/norm.py
torchtitan/models/common/param_init.py
torchtitan/models/common/rope.py
torchtitan/models/common/vision_encoder.py
torchtitan/models/common/vision_encoder_sharding.py
torchtitan/models/common/attention/__init__.py
torchtitan/models/common/attention/attention.py
torchtitan/models/common/attention/cp_attention.py
torchtitan/models/common/attention/cp_gdn.py
torchtitan/models/common/attention/cp_kda.py
torchtitan/models/common/attention/cp_linear_attention.py
torchtitan/models/common/attention/cp_mla.py
torchtitan/models/common/attention/gdn.py
torchtitan/models/common/attention/kda.py
torchtitan/models/common/attention/linear_attention.py
torchtitan/models/common/attention/mla.py
torchtitan/models/common/dist_moe/__init__.py
torchtitan/models/common/dist_moe/mxfp8.py
torchtitan/models/common/dist_moe/routed_experts.py
torchtitan/models/common/dist_moe/runtime.py
torchtitan/models/common/linear/__init__.py
torchtitan/models/common/linear/async_linear.py
torchtitan/models/common/linear/hi_mid_lo_linear.py
torchtitan/models/common/linear/inplace_wgrad_accum.py
torchtitan/models/common/linear/linear.py
torchtitan/models/common/moe/__init__.py
torchtitan/models/common/moe/moe.py
torchtitan/models/common/moe/sharding.py
torchtitan/models/common/moe/state_dict_adapter.py
torchtitan/models/common/moe/token_dispatcher.py
torchtitan/models/deepseek_v3/__init__.py
torchtitan/models/deepseek_v3/attention.py
torchtitan/models/deepseek_v3/flavors.py
torchtitan/models/deepseek_v3/model.py
torchtitan/models/deepseek_v3/moe.py
torchtitan/models/deepseek_v3/mtp.py
torchtitan/models/deepseek_v3/sharding.py
torchtitan/models/deepseek_v3/state_dict_adapter.py
torchtitan/models/deepseek_v4/__init__.py
torchtitan/models/deepseek_v4/attention.py
torchtitan/models/deepseek_v4/compressor.py
torchtitan/models/deepseek_v4/flavors.py
torchtitan/models/deepseek_v4/mhc.py
torchtitan/models/deepseek_v4/model.py
torchtitan/models/deepseek_v4/moe.py
torchtitan/models/deepseek_v4/mtp.py
torchtitan/models/deepseek_v4/sharding.py
torchtitan/models/deepseek_v4/state_dict_adapter.py
torchtitan/models/flux/__init__.py
torchtitan/models/flux/autoencoder.py
torchtitan/models/flux/data.py
torchtitan/models/flux/flavors.py
torchtitan/models/flux/hf_embedder.py
torchtitan/models/flux/layers.py
torchtitan/models/flux/model.py
torchtitan/models/flux/sharding.py
torchtitan/models/flux/state_dict_adapter.py
torchtitan/models/flux/tokenizer.py
torchtitan/models/flux/trainer.py
torchtitan/models/flux/utils.py
torchtitan/models/flux/validator.py
torchtitan/models/flux/inference/infer.py
torchtitan/models/flux/inference/sampling.py
torchtitan/models/gpt_oss/__init__.py
torchtitan/models/gpt_oss/flavors.py
torchtitan/models/gpt_oss/model.py
torchtitan/models/gpt_oss/moe.py
torchtitan/models/gpt_oss/sharding.py
torchtitan/models/gpt_oss/state_dict_adapter.py
torchtitan/models/kimi_k2_7/__init__.py
torchtitan/models/kimi_k2_7/flavors.py
torchtitan/models/kimi_k2_7/model.py
torchtitan/models/kimi_k2_7/qk_clip.py
torchtitan/models/kimi_k2_7/sharding.py
torchtitan/models/kimi_k2_7/state_dict_adapter.py
torchtitan/models/kimi_k2_7/vision_encoder.py
torchtitan/models/kimi_k3/__init__.py
torchtitan/models/kimi_k3/attention.py
torchtitan/models/kimi_k3/flavors.py
torchtitan/models/kimi_k3/model.py
torchtitan/models/kimi_k3/moe.py
torchtitan/models/kimi_k3/sharding.py
torchtitan/models/kimi_k3/state_dict_adapter.py
torchtitan/models/kimi_k3/vision_encoder.py
torchtitan/models/kimi_k3/pipeline_parallel/__init__.py
torchtitan/models/kimi_k3/pipeline_parallel/cache.py
torchtitan/models/kimi_k3/pipeline_parallel/layout.py
torchtitan/models/kimi_k3/pipeline_parallel/stage.py
torchtitan/models/llama3/__init__.py
torchtitan/models/llama3/flavors.py
torchtitan/models/llama3/model.py
torchtitan/models/llama3/state_dict_adapter.py
torchtitan/models/muse_glimmer/__init__.py
torchtitan/models/muse_glimmer/atem.py
torchtitan/models/muse_glimmer/flavors.py
torchtitan/models/muse_glimmer/model.py
torchtitan/models/muse_glimmer/renderer.py
torchtitan/models/muse_glimmer/sharding.py
torchtitan/models/muse_glimmer/state_dict_adapter.py
torchtitan/models/muse_spark/__init__.py
torchtitan/models/muse_spark/audio_encoder.py
torchtitan/models/muse_spark/flavors.py
torchtitan/models/muse_spark/model.py
torchtitan/models/muse_spark/moe.py
torchtitan/models/muse_spark/renderer.py
torchtitan/models/muse_spark/sharding.py
torchtitan/models/muse_spark/state_dict_adapter.py
torchtitan/models/muse_spark/tokenizer.py
torchtitan/models/muse_spark/vision_encoder.py
torchtitan/models/muse_spark/vision_encoder_sharding.py
torchtitan/models/qwen3/__init__.py
torchtitan/models/qwen3/flavors.py
torchtitan/models/qwen3/model.py
torchtitan/models/qwen3/state_dict_adapter.py
torchtitan/models/qwen3_5/__init__.py
torchtitan/models/qwen3_5/flavors.py
torchtitan/models/qwen3_5/model.py
torchtitan/models/qwen3_5/moe.py
torchtitan/models/qwen3_5/rope.py
torchtitan/models/qwen3_5/sharding.py
torchtitan/models/qwen3_5/state_dict_adapter.py
torchtitan/models/qwen3_5/vision_encoder.py
torchtitan/models/qwen3_6/__init__.py
torchtitan/models/qwen3_6/flavors.py
torchtitan/models/qwen3_8/__init__.py
torchtitan/models/qwen3_8/flavors.py
torchtitan/observability/__init__.py
torchtitan/observability/logging.py
torchtitan/observability/metrics.py
torchtitan/observability/profiler.py
torchtitan/observability/sdc_replayer.py
torchtitan/observability/structured_logger/__init__.py
torchtitan/observability/structured_logger/gantt_generator.py
torchtitan/observability/structured_logger/jsonl_handler.py
torchtitan/observability/structured_logger/step_state.py
torchtitan/observability/structured_logger/structured_logging.py
torchtitan/ops/__init__.py
torchtitan/ops/scatter_add.py
torchtitan/quantization/__init__.py
torchtitan/quantization/_fsdp_tensor.py
torchtitan/quantization/utils.py
torchtitan/quantization/mxfp8/__init__.py
torchtitan/quantization/mxfp8/experts.py
torchtitan/quantization/mxfp8/linear.py
torchtitan/quantization/mxfp8/tensor.py
torchtitan/quantization/nvfp4/__init__.py
torchtitan/quantization/nvfp4/experts.py
torchtitan/quantization/nvfp4/linear.py
torchtitan/quantization/nvfp4/tensor.py
torchtitan/quantization/nvfp4/utils.py
torchtitan/rl/__init__.py
torchtitan/rl/_env.py
torchtitan/rl/generate.py
torchtitan/rl/train.py
torchtitan/rl/actors/__init__.py
torchtitan/rl/actors/system.py
torchtitan/rl/actors/monarch/__init__.py
torchtitan/rl/actors/monarch/batcher.py
torchtitan/rl/actors/monarch/generator.py
torchtitan/rl/actors/monarch/provisioning.py
torchtitan/rl/actors/monarch/rdma_microbatch.py
torchtitan/rl/actors/monarch/rollout_worker.py
torchtitan/rl/actors/monarch/router.py
torchtitan/rl/actors/monarch/system.py
torchtitan/rl/actors/monarch/torch_elastic.py
torchtitan/rl/actors/monarch/trainer.py
torchtitan/rl/controller/__init__.py
torchtitan/rl/controller/controller.py
torchtitan/rl/controller/metrics.py
torchtitan/rl/controller/work_buffer.py
torchtitan/rl/examples/__init__.py
torchtitan/rl/examples/alphabet_sort/__init__.py
torchtitan/rl/examples/alphabet_sort/data.py
torchtitan/rl/examples/alphabet_sort/env.py
torchtitan/rl/examples/alphabet_sort/rubric.py
torchtitan/rl/examples/dapo_math/__init__.py
torchtitan/rl/examples/dapo_math/data.py
torchtitan/rl/examples/dapo_math/env.py
torchtitan/rl/examples/dapo_math/math_verify_pool.py
torchtitan/rl/examples/dapo_math/rubric.py
torchtitan/rl/examples/search_r1/__init__.py
torchtitan/rl/examples/search_r1/data.py
torchtitan/rl/examples/search_r1/env.py
torchtitan/rl/examples/search_r1/rubric.py
torchtitan/rl/examples/verifiers/__init__.py
torchtitan/rl/examples/verifiers/data.py
torchtitan/rl/examples/verifiers/env_server.py
torchtitan/rl/examples/verifiers/generation_server.py
torchtitan/rl/examples/verifiers/rollouter.py
torchtitan/rl/examples/verifiers/dapo_math/__init__.py
torchtitan/rl/examples/verifiers/dapo_math/data.py
torchtitan/rl/examples/verifiers/terminal_bench/__init__.py
torchtitan/rl/examples/verifiers/terminal_bench/harness.py
torchtitan/rl/examples/verifiers/terminal_bench/prepare_tmax.py
torchtitan/rl/examples/verifiers/terminal_bench/taskset.py
torchtitan/rl/examples/verifiers/terminal_bench/terminus_harness.py
torchtitan/rl/generator/__init__.py
torchtitan/rl/generator/dispatcher.py
torchtitan/rl/generator/engine.py
torchtitan/rl/generator/engine_loop.py
torchtitan/rl/generator/generator.py
torchtitan/rl/generator/parallelism.py
torchtitan/rl/generator/types.py
torchtitan/rl/generator/vllm/__init__.py
torchtitan/rl/generator/vllm/attention.py
torchtitan/rl/generator/vllm/batch_invariance.py
torchtitan/rl/generator/vllm/config.py
torchtitan/rl/generator/vllm/engine.py
torchtitan/rl/generator/vllm/gdn.py
torchtitan/rl/generator/vllm/kda.py
torchtitan/rl/generator/vllm/linear_attention.py
torchtitan/rl/generator/vllm/model_wrapper.py
torchtitan/rl/generator/vllm/registry.py
torchtitan/rl/generator/vllm/session_kv_holder.py
torchtitan/rl/generator/vllm/stat_logger.py
torchtitan/rl/generator/vllm/worker.py
torchtitan/rl/observability/__init__.py
torchtitan/rl/observability/metrics/__init__.py
torchtitan/rl/observability/metrics/console.py
torchtitan/rl/observability/metrics/processor.py
torchtitan/rl/observability/metrics/types.py
torchtitan/rl/rollout/__init__.py
torchtitan/rl/rollout/advantage.py
torchtitan/rl/rollout/data.py
torchtitan/rl/rollout/message_env.py
torchtitan/rl/rollout/recorder.py
torchtitan/rl/rollout/rollouter.py
torchtitan/rl/rollout/rubric.py
torchtitan/rl/rollout/token_env.py
torchtitan/rl/rollout/types.py
torchtitan/rl/routing/__init__.py
torchtitan/rl/routing/inter_generator.py
torchtitan/rl/routing/intra_generator.py
torchtitan/rl/routing/strategies.py
torchtitan/rl/routing/types.py
torchtitan/rl/trainer/__init__.py
torchtitan/rl/trainer/batcher.py
torchtitan/rl/trainer/controller_state.py
torchtitan/rl/trainer/rdma_microbatch.py
torchtitan/rl/trainer/trainer.py
torchtitan/rl/trainer/training_sample_builder.py
torchtitan/rl/trainer/losses/__init__.py
torchtitan/rl/trainer/losses/dapo.py
torchtitan/rl/trainer/losses/grpo.py
torchtitan/tools/filesystem.py
torchtitan/tools/garbage_collector.py
torchtitan/tools/utils.py
torchtitan_recipes/__init__.py
torchtitan_recipes/data.py
torchtitan_recipes/graph_trainer/__init__.py
torchtitan_recipes/graph_trainer/deepseek_v3.py
torchtitan_recipes/graph_trainer/llama3.py
torchtitan_recipes/graph_trainer/qwen3.py
torchtitan_recipes/models/__init__.py
torchtitan_recipes/models/deepseek_v3.py
torchtitan_recipes/models/flux.py
torchtitan_recipes/models/llama3.py
torchtitan_recipes/models/muse_glimmer.py
torchtitan_recipes/models/muse_spark.py
torchtitan_recipes/models/qwen3.py
torchtitan_recipes/overrides/__init__.py
torchtitan_recipes/overrides/fused_mla.py
torchtitan_recipes/overrides/fused_swiglu.py
torchtitan_recipes/overrides/kimi_attention_residual.py
torchtitan_recipes/rl/__init__.py
torchtitan_recipes/rl/alphabet_sort.py
torchtitan_recipes/rl/dapo_math.py
torchtitan_recipes/rl/model_configs.py
torchtitan_recipes/rl/search_r1.py
torchtitan_recipes/rl/verifiers_dapo_math.py
torchtitan_recipes/rl/verifiers_terminal_bench.py
torchtitan_recipes/tests/__init__.py
torchtitan_recipes/tests/transformers_modeling_backend.py
torchtitan_recipes/tests/datasets/__init__.py
torchtitan_recipes/tests/datasets/synthetic_multimodal.py
torchtitan_recipes/tests/graph_trainer/__init__.py
torchtitan_recipes/tests/graph_trainer/b200.py
torchtitan_recipes/tests/graph_trainer/deepseek_v3.py
torchtitan_recipes/tests/graph_trainer/llama3.py
torchtitan_recipes/tests/graph_trainer/muse_glimmer.py
torchtitan_recipes/tests/graph_trainer/numerics.py
torchtitan_recipes/tests/graph_trainer/precompile.py
torchtitan_recipes/tests/graph_trainer/qwen3.py
torchtitan_recipes/tests/models/__init__.py
torchtitan_recipes/tests/models/deepseek_v3.py
torchtitan_recipes/tests/models/deepseek_v4.py
torchtitan_recipes/tests/models/flux.py
torchtitan_recipes/tests/models/gpt_oss.py
torchtitan_recipes/tests/models/kimi_k2_7.py
torchtitan_recipes/tests/models/kimi_k3.py
torchtitan_recipes/tests/models/llama3.py
torchtitan_recipes/tests/models/multimodal.py
torchtitan_recipes/tests/models/muse_glimmer.py
torchtitan_recipes/tests/models/muse_spark.py
torchtitan_recipes/tests/models/qwen3.py
torchtitan_recipes/tests/models/qwen3_5.py
torchtitan_recipes/tests/models/qwen3_6.py
torchtitan_recipes/tests/models/qwen3_8.py
torchtitan_recipes/tests/rl/__init__.py
torchtitan_recipes/tests/rl/alphabet_sort.py
torchtitan_recipes/tests/torchft/__init__.py
torchtitan_recipes/tests/torchft/llama3.py
torchtitan_recipes/torchft/__init__.py
torchtitan_recipes/torchft/llama3.py