diff --git a/.gitattributes b/.gitattributes index 52f8f26867ff0b240eb4dc16b1754a9f568e8559..b926ef9a3f570aa606e26dea957e31cccd3bb8fd 100644 --- a/.gitattributes +++ b/.gitattributes @@ -492,3 +492,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text 2026.04.23/15.32.08_train_llm_diffusion_unet_lowdim_box-close-v2/wandb/run-20260423_153304-it8p5fpg/run-it8p5fpg.wandb filter=lfs diff=lfs merge=lfs -text 2026.04.24/19.32.45_train_llm_diffusion_unet_lowdim_box-close-v2/logs.json.txt filter=lfs diff=lfs merge=lfs -text 2026.04.24/19.32.45_train_llm_diffusion_unet_lowdim_box-close-v2/wandb/run-20260424_193342-4hpop9xi/run-4hpop9xi.wandb filter=lfs diff=lfs merge=lfs -text +2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/logs.json.txt filter=lfs diff=lfs merge=lfs -text +2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/wandb/run-20260425_035339-ivkyq15y/run-ivkyq15y.wandb filter=lfs diff=lfs merge=lfs -text +2026.04.25/03.52.11_train_llm_diffusion_unet_lowdim_sweep-v2/logs.json.txt filter=lfs diff=lfs merge=lfs -text +2026.04.25/03.52.11_train_llm_diffusion_unet_lowdim_sweep-v2/wandb/run-20260425_035342-iafad9fe/run-iafad9fe.wandb filter=lfs diff=lfs merge=lfs -text +2026.04.25/22.26.21_train_llm_diffusion_unet_lowdim_box-close-v2/logs.json.txt filter=lfs diff=lfs merge=lfs -text +2026.04.25/22.26.21_train_llm_diffusion_unet_lowdim_box-close-v2/wandb/run-20260425_222718-n1fpkzkm/run-n1fpkzkm.wandb filter=lfs diff=lfs merge=lfs -text diff --git a/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..38dd678db12ad1be81cfa372fd9b1504041c2bcb --- /dev/null +++ b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml @@ -0,0 +1,193 @@ +name: train_llm_diffusion_unet_lowdim +_target_: llmbc.workspace.train_llm_diffusion_unet_lowdim_workspace.TrainLLMDiffusionUnetLowdimWorkspace +obs_dim: ${task.obs_dim} +action_dim: ${task.action_dim} +task_name: ${task.name} +exp_name: default +model_name: ${llm.name} +horizon: 1 +n_obs_steps: 1 +n_action_steps: 1 +n_latency_steps: 0 +past_action_visible: false +keypoint_visible_rate: 1.0 +obs_as_local_cond: false +obs_as_global_cond: true +pred_action_steps_only: false +llm_orig_expert_feedback: true +llm_do_sample: false +policy: + _target_: llmbc.policy.llm_diffusion_unet_lowdim_policy.LLMDiffusionUnetLowdimPolicy + model: + _target_: llmbc.model.diffusion.conditional_unet1d.ConditionalUnet1D + input_dim: '${eval: ${task.action_dim} if ${obs_as_local_cond} or ${obs_as_global_cond} + else ${task.obs_dim} + ${task.action_dim}}' + local_cond_dim: '${eval: ${task.obs_dim} if ${obs_as_local_cond} else None}' + global_cond_dim: '${eval: ${task.obs_dim}*${n_obs_steps} if ${obs_as_global_cond} + else None}' + diffusion_step_embed_dim: 256 + down_dims: + - 256 + - 512 + - 1024 + kernel_size: 5 + n_groups: 8 + cond_predict_scale: true + noise_scheduler: + _target_: diffusers.schedulers.scheduling_ddpm.DDPMScheduler + num_train_timesteps: 100 + beta_start: 0.0001 + beta_end: 0.02 + beta_schedule: squaredcos_cap_v2 + variance_type: fixed_small + clip_sample: true + prediction_type: epsilon + horizon: ${horizon} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + n_obs_steps: ${n_obs_steps} + num_inference_steps: 100 + obs_as_local_cond: ${obs_as_local_cond} + obs_as_global_cond: ${obs_as_global_cond} + pred_action_steps_only: ${pred_action_steps_only} + oa_step_convention: true + llm_discriminator: + _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator + task_id: ${task_name} + llm_translator: + _target_: llmbc.translator.llm_translator.LLMTranslator + cfg: ${llm} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + horizon: ${horizon} + n_obs_steps: ${n_obs_steps} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + loss_dp_weight: 1.0 + loss_llm_weight: 0.001 + normalize_llm_loss: true + reweight_llm_loss: true +ema: + _target_: llmbc.model.diffusion.ema_model.EMAModel + update_after_step: 0 + inv_gamma: 1.0 + power: 0.75 + min_value: 0.0 + max_value: 0.9999 +dataloader: + batch_size: 16 + num_workers: 0 + shuffle: true + pin_memory: false + persistent_workers: false +val_dataloader: + batch_size: 16 + num_workers: 0 + shuffle: false + pin_memory: false + persistent_workers: false +optimizer: + _target_: torch.optim.AdamW + lr: 0.0001 + betas: + - 0.95 + - 0.999 + eps: 1.0e-08 + weight_decay: 1.0e-06 +training: + device: cuda:0 + seed: 42 + debug: false + resume: true + lr_scheduler: cosine + lr_warmup_steps: 500 + num_epochs: 1001 + gradient_accumulate_every: 16 + use_ema: true + rollout_every: 5 + checkpoint_every: 5 + val_every: 1 + sample_every: 5 + max_train_steps: null + max_val_steps: null + tqdm_interval_sec: 1.0 +logging: + project: ${task.name}-training-mix + resume: true + mode: online + name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} + tags: + - ${name} + - ${task_name} + - ${exp_name} + id: null + group: null +checkpoint: + topk: + monitor_key: test_success_rate + mode: max + k: 5 + format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt + save_last_ckpt: true + save_last_snapshot: false +multi_run: + run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} +task: + name: sweep-v2 + obs_dim: 9 + action_dim: 4 + env_runner: + _target_: llmbc.env_runner.metaworld_lowdim_runner.MetaworldLowdimRunner + env_name: llf-metaworld-sweep-v2 + n_train: 10 + n_test: 50 + n_envs: 10 + max_steps: 30 + n_obs_steps: ${n_obs_steps} + n_action_steps: ${n_action_steps} + instruction_type: b + feedback_type: + - hp + - hn + - fp + visual: false + discount: 0.9 + dataset: + _target_: llmbc.dataset.metaworld_lowdim_dataset.MetaworldLowdimDataset + data_path: datasets/sweep-v2-mix(csp3).pt + data_path2: datasets/sweep-v2.pt + horizon: ${horizon} + pad_before: ${eval:'${n_obs_steps}-1'} + pad_after: ${eval:'${n_action_steps}-1'} + obs_eef_target: true + use_manual_normalizer: false + val_ratio: 0.1 + dummy_normalizer: false + instructor: + _target_: llmbc.translator.instructor.metaworld_instructor.sweep_v2_instructor.SweepV2Instructor +llm: + name: HuggingFaceTB/SmolLM2-135M-Instruct + model_name: SmolLM2-135M-Instruct + config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig + causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM + use_quantization: false + use_joint_mlp_projector: true + llm_mode: ete-finetuned + finetune_mode: orig + checkpoint: data/outputs/2025.11.01/23.03.01_train_llm_lowdim_sweep-v2/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-sweep-v2/checkpoint-5690 + max_length: 100 + lora_config: + r: 32 + lora_alpha: 64 + lora_dropout: 0.05 + bias: none + task_type: CAUSAL_LM + prompter: + _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter + use_joint_mlp_projector: true + hydra: + job: + override_dirname: ${model_name} + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name} diff --git a/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b2f8461014313d275ec77a7eb1c29a3aa8b7aa3e --- /dev/null +++ b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml @@ -0,0 +1,155 @@ +hydra: + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + sweep: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - training.seed=42 + job: + name: train + chdir: null + override_dirname: training.seed=42 + id: ??? + num: ??? + config_name: llmdp_sweep-v2_mix(csp3).yaml + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.2.0 + version_base: '1.2' + cwd: /workspace/LLM-BC + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/LLM-BC/config/mix_data + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/LLM-BC/data/outputs/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2 + choices: + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b8fcfadfbd1253a3af726bb8f893b41de23943fc --- /dev/null +++ b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml @@ -0,0 +1 @@ +- training.seed=42 diff --git a/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/train.log b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..48f28d1f56355ee1b8ff80ed70d9e7c2bd422e87 --- /dev/null +++ b/2026.04.25/03.47.13_train_llm_diffusion_unet_lowdim_sweep-v2/train.log @@ -0,0 +1,7 @@ +[2026-04-25 03:47:15,107][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. +[2026-04-25 03:47:15,107][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16. +[2026-04-25 03:47:15,107][numexpr.utils][INFO] - NumExpr defaulting to 16 threads. +[2026-04-25 03:47:17,304][datasets][INFO] - PyTorch version 2.2.2 available. +[2026-04-25 03:47:17,305][datasets][INFO] - TensorFlow version 2.15.1 available. +[2026-04-25 03:47:17,305][datasets][INFO] - JAX version 0.4.30 available. +[2026-04-25 03:47:19,027][llmbc.model.diffusion.conditional_unet1d][INFO] - number of parameters: 6.501479e+07 diff --git a/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..74f2c5ba7e7720ee3c0180158d32847b48c181da --- /dev/null +++ b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml @@ -0,0 +1,193 @@ +name: train_llm_diffusion_unet_lowdim +_target_: llmbc.workspace.train_llm_diffusion_unet_lowdim_workspace.TrainLLMDiffusionUnetLowdimWorkspace +obs_dim: ${task.obs_dim} +action_dim: ${task.action_dim} +task_name: ${task.name} +exp_name: default +model_name: ${llm.name} +horizon: 1 +n_obs_steps: 1 +n_action_steps: 1 +n_latency_steps: 0 +past_action_visible: false +keypoint_visible_rate: 1.0 +obs_as_local_cond: false +obs_as_global_cond: true +pred_action_steps_only: false +llm_orig_expert_feedback: true +llm_do_sample: false +policy: + _target_: llmbc.policy.llm_diffusion_unet_lowdim_policy.LLMDiffusionUnetLowdimPolicy + model: + _target_: llmbc.model.diffusion.conditional_unet1d.ConditionalUnet1D + input_dim: '${eval: ${task.action_dim} if ${obs_as_local_cond} or ${obs_as_global_cond} + else ${task.obs_dim} + ${task.action_dim}}' + local_cond_dim: '${eval: ${task.obs_dim} if ${obs_as_local_cond} else None}' + global_cond_dim: '${eval: ${task.obs_dim}*${n_obs_steps} if ${obs_as_global_cond} + else None}' + diffusion_step_embed_dim: 256 + down_dims: + - 256 + - 512 + - 1024 + kernel_size: 5 + n_groups: 8 + cond_predict_scale: true + noise_scheduler: + _target_: diffusers.schedulers.scheduling_ddpm.DDPMScheduler + num_train_timesteps: 100 + beta_start: 0.0001 + beta_end: 0.02 + beta_schedule: squaredcos_cap_v2 + variance_type: fixed_small + clip_sample: true + prediction_type: epsilon + horizon: ${horizon} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + n_obs_steps: ${n_obs_steps} + num_inference_steps: 100 + obs_as_local_cond: ${obs_as_local_cond} + obs_as_global_cond: ${obs_as_global_cond} + pred_action_steps_only: ${pred_action_steps_only} + oa_step_convention: true + llm_discriminator: + _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator + task_id: ${task_name} + llm_translator: + _target_: llmbc.translator.llm_translator.LLMTranslator + cfg: ${llm} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + horizon: ${horizon} + n_obs_steps: ${n_obs_steps} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + loss_dp_weight: 1.0 + loss_llm_weight: 0.001 + normalize_llm_loss: true + reweight_llm_loss: true +ema: + _target_: llmbc.model.diffusion.ema_model.EMAModel + update_after_step: 0 + inv_gamma: 1.0 + power: 0.75 + min_value: 0.0 + max_value: 0.9999 +dataloader: + batch_size: 16 + num_workers: 0 + shuffle: true + pin_memory: false + persistent_workers: false +val_dataloader: + batch_size: 16 + num_workers: 0 + shuffle: false + pin_memory: false + persistent_workers: false +optimizer: + _target_: torch.optim.AdamW + lr: 0.0001 + betas: + - 0.95 + - 0.999 + eps: 1.0e-08 + weight_decay: 1.0e-06 +training: + device: cuda:0 + seed: 43 + debug: false + resume: true + lr_scheduler: cosine + lr_warmup_steps: 500 + num_epochs: 1001 + gradient_accumulate_every: 16 + use_ema: true + rollout_every: 5 + checkpoint_every: 5 + val_every: 1 + sample_every: 5 + max_train_steps: null + max_val_steps: null + tqdm_interval_sec: 1.0 +logging: + project: ${task.name}-training-mix + resume: true + mode: online + name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} + tags: + - ${name} + - ${task_name} + - ${exp_name} + id: null + group: null +checkpoint: + topk: + monitor_key: test_success_rate + mode: max + k: 5 + format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt + save_last_ckpt: true + save_last_snapshot: false +multi_run: + run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} +task: + name: sweep-v2 + obs_dim: 9 + action_dim: 4 + env_runner: + _target_: llmbc.env_runner.metaworld_lowdim_runner.MetaworldLowdimRunner + env_name: llf-metaworld-sweep-v2 + n_train: 10 + n_test: 50 + n_envs: 10 + max_steps: 30 + n_obs_steps: ${n_obs_steps} + n_action_steps: ${n_action_steps} + instruction_type: b + feedback_type: + - hp + - hn + - fp + visual: false + discount: 0.9 + dataset: + _target_: llmbc.dataset.metaworld_lowdim_dataset.MetaworldLowdimDataset + data_path: datasets/sweep-v2-mix(csp3).pt + data_path2: datasets/sweep-v2.pt + horizon: ${horizon} + pad_before: ${eval:'${n_obs_steps}-1'} + pad_after: ${eval:'${n_action_steps}-1'} + obs_eef_target: true + use_manual_normalizer: false + val_ratio: 0.1 + dummy_normalizer: false + instructor: + _target_: llmbc.translator.instructor.metaworld_instructor.sweep_v2_instructor.SweepV2Instructor +llm: + name: HuggingFaceTB/SmolLM2-135M-Instruct + model_name: SmolLM2-135M-Instruct + config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig + causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM + use_quantization: false + use_joint_mlp_projector: true + llm_mode: ete-finetuned + finetune_mode: orig + checkpoint: data/outputs/2025.11.01/23.03.01_train_llm_lowdim_sweep-v2/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-sweep-v2/checkpoint-5690 + max_length: 100 + lora_config: + r: 32 + lora_alpha: 64 + lora_dropout: 0.05 + bias: none + task_type: CAUSAL_LM + prompter: + _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter + use_joint_mlp_projector: true + hydra: + job: + override_dirname: ${model_name} + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name} diff --git a/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b15a9678e036c4b1d21c7e0481006fed67de0911 --- /dev/null +++ b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml @@ -0,0 +1,155 @@ +hydra: + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + sweep: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - training.seed=43 + job: + name: train + chdir: null + override_dirname: training.seed=43 + id: ??? + num: ??? + config_name: llmdp_sweep-v2_mix(csp3).yaml + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.2.0 + version_base: '1.2' + cwd: /workspace/LLM-BC + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/LLM-BC/config/mix_data + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/LLM-BC/data/outputs/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2 + choices: + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ebf61aeca24956094799b4a1cab9370df040027f --- /dev/null +++ b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml @@ -0,0 +1 @@ +- training.seed=43 diff --git a/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/train.log b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..cdf32cb863a831d8c860990d5f15dafb1bf547f3 --- /dev/null +++ b/2026.04.25/03.47.19_train_llm_diffusion_unet_lowdim_sweep-v2/train.log @@ -0,0 +1,7 @@ +[2026-04-25 03:47:21,387][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. +[2026-04-25 03:47:21,387][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16. +[2026-04-25 03:47:21,387][numexpr.utils][INFO] - NumExpr defaulting to 16 threads. +[2026-04-25 03:47:23,695][datasets][INFO] - PyTorch version 2.2.2 available. +[2026-04-25 03:47:23,696][datasets][INFO] - TensorFlow version 2.15.1 available. +[2026-04-25 03:47:23,696][datasets][INFO] - JAX version 0.4.30 available. +[2026-04-25 03:47:25,356][llmbc.model.diffusion.conditional_unet1d][INFO] - number of parameters: 6.501479e+07 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..38dd678db12ad1be81cfa372fd9b1504041c2bcb --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/config.yaml @@ -0,0 +1,193 @@ +name: train_llm_diffusion_unet_lowdim +_target_: llmbc.workspace.train_llm_diffusion_unet_lowdim_workspace.TrainLLMDiffusionUnetLowdimWorkspace +obs_dim: ${task.obs_dim} +action_dim: ${task.action_dim} +task_name: ${task.name} +exp_name: default +model_name: ${llm.name} +horizon: 1 +n_obs_steps: 1 +n_action_steps: 1 +n_latency_steps: 0 +past_action_visible: false +keypoint_visible_rate: 1.0 +obs_as_local_cond: false +obs_as_global_cond: true +pred_action_steps_only: false +llm_orig_expert_feedback: true +llm_do_sample: false +policy: + _target_: llmbc.policy.llm_diffusion_unet_lowdim_policy.LLMDiffusionUnetLowdimPolicy + model: + _target_: llmbc.model.diffusion.conditional_unet1d.ConditionalUnet1D + input_dim: '${eval: ${task.action_dim} if ${obs_as_local_cond} or ${obs_as_global_cond} + else ${task.obs_dim} + ${task.action_dim}}' + local_cond_dim: '${eval: ${task.obs_dim} if ${obs_as_local_cond} else None}' + global_cond_dim: '${eval: ${task.obs_dim}*${n_obs_steps} if ${obs_as_global_cond} + else None}' + diffusion_step_embed_dim: 256 + down_dims: + - 256 + - 512 + - 1024 + kernel_size: 5 + n_groups: 8 + cond_predict_scale: true + noise_scheduler: + _target_: diffusers.schedulers.scheduling_ddpm.DDPMScheduler + num_train_timesteps: 100 + beta_start: 0.0001 + beta_end: 0.02 + beta_schedule: squaredcos_cap_v2 + variance_type: fixed_small + clip_sample: true + prediction_type: epsilon + horizon: ${horizon} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + n_obs_steps: ${n_obs_steps} + num_inference_steps: 100 + obs_as_local_cond: ${obs_as_local_cond} + obs_as_global_cond: ${obs_as_global_cond} + pred_action_steps_only: ${pred_action_steps_only} + oa_step_convention: true + llm_discriminator: + _target_: llmbc.discriminator.llm_ce_discriminator.LLMCEDiscriminator + task_id: ${task_name} + llm_translator: + _target_: llmbc.translator.llm_translator.LLMTranslator + cfg: ${llm} + obs_dim: ${obs_dim} + action_dim: ${action_dim} + horizon: ${horizon} + n_obs_steps: ${n_obs_steps} + n_action_steps: ${eval:'${n_action_steps}+${n_latency_steps}'} + loss_dp_weight: 1.0 + loss_llm_weight: 0.001 + normalize_llm_loss: true + reweight_llm_loss: true +ema: + _target_: llmbc.model.diffusion.ema_model.EMAModel + update_after_step: 0 + inv_gamma: 1.0 + power: 0.75 + min_value: 0.0 + max_value: 0.9999 +dataloader: + batch_size: 16 + num_workers: 0 + shuffle: true + pin_memory: false + persistent_workers: false +val_dataloader: + batch_size: 16 + num_workers: 0 + shuffle: false + pin_memory: false + persistent_workers: false +optimizer: + _target_: torch.optim.AdamW + lr: 0.0001 + betas: + - 0.95 + - 0.999 + eps: 1.0e-08 + weight_decay: 1.0e-06 +training: + device: cuda:0 + seed: 42 + debug: false + resume: true + lr_scheduler: cosine + lr_warmup_steps: 500 + num_epochs: 1001 + gradient_accumulate_every: 16 + use_ema: true + rollout_every: 5 + checkpoint_every: 5 + val_every: 1 + sample_every: 5 + max_train_steps: null + max_val_steps: null + tqdm_interval_sec: 1.0 +logging: + project: ${task.name}-training-mix + resume: true + mode: online + name: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} + tags: + - ${name} + - ${task_name} + - ${exp_name} + id: null + group: null +checkpoint: + topk: + monitor_key: test_success_rate + mode: max + k: 5 + format_str: epoch={epoch:04d}-test_success_rate={test_success_rate:.3f}.ckpt + save_last_ckpt: true + save_last_snapshot: false +multi_run: + run_dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + wandb_name_base: ${now:%Y.%m.%d-%H.%M.%S}_${name}_${task_name} +task: + name: sweep-v2 + obs_dim: 9 + action_dim: 4 + env_runner: + _target_: llmbc.env_runner.metaworld_lowdim_runner.MetaworldLowdimRunner + env_name: llf-metaworld-sweep-v2 + n_train: 10 + n_test: 50 + n_envs: 10 + max_steps: 30 + n_obs_steps: ${n_obs_steps} + n_action_steps: ${n_action_steps} + instruction_type: b + feedback_type: + - hp + - hn + - fp + visual: false + discount: 0.9 + dataset: + _target_: llmbc.dataset.metaworld_lowdim_dataset.MetaworldLowdimDataset + data_path: datasets/sweep-v2-mix(csp3).pt + data_path2: datasets/sweep-v2.pt + horizon: ${horizon} + pad_before: ${eval:'${n_obs_steps}-1'} + pad_after: ${eval:'${n_action_steps}-1'} + obs_eef_target: true + use_manual_normalizer: false + val_ratio: 0.1 + dummy_normalizer: false + instructor: + _target_: llmbc.translator.instructor.metaworld_instructor.sweep_v2_instructor.SweepV2Instructor +llm: + name: HuggingFaceTB/SmolLM2-135M-Instruct + model_name: SmolLM2-135M-Instruct + config_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaConfig + causal_lm_target: llmbc.model.llm.llama_lowdim_model.LowdimLlamaForCausalLM + use_quantization: false + use_joint_mlp_projector: true + llm_mode: ete-finetuned + finetune_mode: orig + checkpoint: data/outputs/2025.11.01/23.03.01_train_llm_lowdim_sweep-v2/HuggingFaceTB/SmolLM2-135M-Instruct-finetuned-sweep-v2/checkpoint-5690 + max_length: 100 + lora_config: + r: 32 + lora_alpha: 64 + lora_dropout: 0.05 + bias: none + task_type: CAUSAL_LM + prompter: + _target_: llmbc.translator.prompter.smollm2_prompter.SmolLM2Prompter + use_joint_mlp_projector: true + hydra: + job: + override_dirname: ${model_name} + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${model_name} diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b1c215658687eddf194247321b28289eafcf433b --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/hydra.yaml @@ -0,0 +1,155 @@ +hydra: + run: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + sweep: + dir: data/outputs/${now:%Y.%m.%d}/${now:%H.%M.%S}_${name}_${task_name} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - training.seed=42 + job: + name: train + chdir: null + override_dirname: training.seed=42 + id: ??? + num: ??? + config_name: llmdp_sweep-v2_mix(csp3).yaml + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.2.0 + version_base: '1.2' + cwd: /workspace/LLM-BC + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/LLM-BC/config/mix_data + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/LLM-BC/data/outputs/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2 + choices: + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b8fcfadfbd1253a3af726bb8f893b41de23943fc --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/.hydra/overrides.yaml @@ -0,0 +1 @@ +- training.seed=42 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0180-test_success_rate=0.740.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0180-test_success_rate=0.740.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..3ecc0b082604502b9da49a80f3d604f6fbe9e9a6 --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0180-test_success_rate=0.740.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9374a1bd4fa15f7d2831b4083fe9ace1a39357b51c6c1577e4cdb4af4132b707 +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0265-test_success_rate=0.680.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0265-test_success_rate=0.680.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..060fdd3d1147f8b0cf43104336bc0621cbea5adf --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0265-test_success_rate=0.680.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:416fe9b5412b07a3c2bb3b02f63d7370858921112ee5e143d27859bb6452f253 +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0410-test_success_rate=0.680.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0410-test_success_rate=0.680.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..c69ebe3b856c8039ebfdf8492d4c3487bb944e63 --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0410-test_success_rate=0.680.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c91a2faee47005f5e07b7306020afe42359a262fafbfb3418136cdd066725b77 +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0455-test_success_rate=0.780.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0455-test_success_rate=0.780.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..570b06fe8fe088df76f620e67e3011ca40f67815 --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0455-test_success_rate=0.780.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d535b2cf88fe4d3ceeb3cbe86c20d989a9e4c1c7a0681cb8876055709def17a +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0515-test_success_rate=0.680.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0515-test_success_rate=0.680.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..a2c38f3095d191aed095e10a7d7dea7a0d11cbed --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/epoch=0515-test_success_rate=0.680.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:623eb4c2f6bb2e3ab206cee4358793169e412fa7c610c6c1326555f3f77c432d +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/latest.ckpt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/latest.ckpt new file mode 100644 index 0000000000000000000000000000000000000000..c89baa8142cf6f41437ac776e30a59dccbf8aacf --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/checkpoints/latest.ckpt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d0814a9d001e4da156135f8654859113aae6d0f3694a623b9f2711a4ad36e3c0 +size 1040498994 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/logs.json.txt b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/logs.json.txt new file mode 100644 index 0000000000000000000000000000000000000000..af079be8fdf1042f686387f04ec781af98510a14 --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/logs.json.txt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22e8eca293b2a1177ac050839d717267ac006bb060e1c8cf456e518d0d76ccee +size 120440567 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/train.log b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..018477583929dcec55b24989c0bd773687515e36 --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/train.log @@ -0,0 +1,14 @@ +[2026-04-25 03:52:04,619][numexpr.utils][INFO] - Note: detected 144 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. +[2026-04-25 03:52:04,619][numexpr.utils][INFO] - Note: NumExpr detected 144 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 16. +[2026-04-25 03:52:04,619][numexpr.utils][INFO] - NumExpr defaulting to 16 threads. +[2026-04-25 03:52:06,853][datasets][INFO] - PyTorch version 2.2.2 available. +[2026-04-25 03:52:06,854][datasets][INFO] - TensorFlow version 2.15.1 available. +[2026-04-25 03:52:06,855][datasets][INFO] - JAX version 0.4.30 available. +[2026-04-25 03:52:08,560][llmbc.model.diffusion.conditional_unet1d][INFO] - number of parameters: 6.501479e+07 +[2026-04-25 03:52:12,545][root][INFO] - running build_ext +[2026-04-25 03:52:12,547][root][INFO] - building 'mujoco_py.cymj' extension +[2026-04-25 03:52:12,547][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o -fopenmp -w +[2026-04-25 03:52:55,118][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -O2 -Wall -fPIC -O2 -isystem /venv/llm-bc/include -I/venv/llm-bc/include -fPIC -O2 -isystem /venv/llm-bc/include -fPIC -I/venv/llm-bc/lib/python3.9/site-packages/mujoco_py -I/root/.mujoco/mujoco210/include -I/venv/llm-bc/lib/python3.9/site-packages/numpy/core/include -I/venv/llm-bc/include/python3.9 -c /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.c -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -fopenmp -w +[2026-04-25 03:52:55,248][root][INFO] - gcc -pthread -B /venv/llm-bc/compiler_compat -shared -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib -L/venv/llm-bc/lib -Wl,-rpath,/venv/llm-bc/lib -Wl,-rpath-link,/venv/llm-bc/lib -L/venv/llm-bc/lib /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/cymj.o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/temp.linux-x86_64-cpython-39/venv/llm-bc/lib/python3.9/site-packages/mujoco_py/gl/osmesashim.o -L/root/.mujoco/mujoco210/bin -Wl,--enable-new-dtags,-rpath,/root/.mujoco/mujoco210/bin -lmujoco210 -lglewosmesa -lOSMesa -lGL -o /venv/llm-bc/lib/python3.9/site-packages/mujoco_py/generated/_pyxbld_2.1.2.14_39_linuxcpuextensionbuilder/lib.linux-x86_64-cpython-39/mujoco_py/cymj.cpython-39-x86_64-linux-gnu.so -fopenmp +[2026-04-25 03:53:38,977][absl][INFO] - MUJOCO_GL=osmesa, attempting to import specified OpenGL backend. +[2026-04-25 03:53:38,987][absl][INFO] - MuJoCo library version is: 2.3.7 diff --git a/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/wandb/run-20260425_035339-ivkyq15y/files/output.log b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/wandb/run-20260425_035339-ivkyq15y/files/output.log new file mode 100644 index 0000000000000000000000000000000000000000..2c9f0b81cf443a02358d2d10ecd400ea6b1711ce --- /dev/null +++ b/2026.04.25/03.52.02_train_llm_diffusion_unet_lowdim_sweep-v2/wandb/run-20260425_035339-ivkyq15y/files/output.log @@ -0,0 +1,3 @@ +Eval MetaworldLowdimRunner 1/6: 0%| | 0/30 [00:00