Upload 6 files

hierarchos.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f34ba416829b300faab486d51659a40a12a907d8fcdefbf14d2ac7d90dabc524
+size 935386573

hierarchos_config.json

@@ -0,0 +1,87 @@
+{
+  "mode": "train",
+  "train": true,
+  "hf_dataset": "netcat420/Experiment_0.1",
+  "hf_dataset_config": null,
+  "hf_dataset_split": "train",
+  "text_column": null,
+  "prompt_column": "instruction",
+  "completion_column": "output",
+  "model_path": null,
+  "out_dir": "./chatHRM",
+  "tokenizer_path": null,
+  "resume_from_ckpt": null,
+  "shadow_model_path": null,
+  "pre_chunked_dataset": false,
+  "pre_pt_dataset": false,
+  "context_dim": 450,
+  "persistent_dim": 128,
+  "ltm_slots": 1024,
+  "ltm_key_dim": 128,
+  "ltm_val_dim": 128,
+  "h_hidden": 450,
+  "l_hidden": 450,
+  "h_stride": 4,
+  "max_h_steps": 5,
+  "max_l_steps": 5,
+  "ltm_topk": 4,
+  "max_length": 10016,
+  "auto_max_length": true,
+  "epochs": 10,
+  "batch_size": 8,
+  "accumulation_steps": 1,
+  "starting_lr": 0.0001,
+  "min_lr": 9e-08,
+  "disable_lr_schedule": false,
+  "ltm_lr": 0.001,
+  "kayla": false,
+  "lora_r": 8,
+  "lora_alpha": 16,
+  "grad_clip": 1.0,
+  "ponder_loss_weight": 0.001,
+  "commitment_loss_weight": 0.5,
+  "commitment_threshold": 0.05,
+  "l_conv_atol": 0.0001,
+  "detach_every_n_steps": 32,
+  "h_halt_thresh": 0.9,
+  "encourage_thinking": false,
+  "adaptive_ponder": false,
+  "ponder_target_scale": 0.5,
+  "reset_halt_bias": null,
+  "override_scheduling": false,
+  "persist_state": false,
+  "training_chunk_size": 128,
+  "save_steps": 0,
+  "num_workers": 0,
+  "amp": true,
+  "compile": true,
+  "force_compile": true,
+  "eval_tasks": null,
+  "eval_every_epoch": 1,
+  "eval_batch_size": 1,
+  "eval_limit": null,
+  "eval_steps": null,
+  "temperature": 0.7,
+  "top_k": 40,
+  "top_p": 0.9,
+  "repetition_penalty": 1.2,
+  "max_new_tokens": 512,
+  "device": null,
+  "threads": 24,
+  "enable_quantized_learning": false,
+  "ltm_lora_path": null,
+  "static_ltm_lr": false,
+  "ltm_schedule_steps": 100,
+  "ltm_schedule_min_lr": 1e-05,
+  "finetune_unlock_percent": null,
+  "gradient_checkpointing": false,
+  "passive_learning": true,
+  "passive_lr": 5e-06,
+  "surprise_threshold": 1.0,
+  "ckpt_input": null,
+  "inf_output": null,
+  "ckpt_tok_path": null,
+  "amp_dtype": "bfloat16",
+  "vocab_size": 50257,
+  "completed_epoch": 28
+}

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "is_local": false,
+  "model_max_length": 10016,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}