add pythia-70m config

config.json

@@ -1,51 +1,24 @@
 {
-  "seed": 0,
-  "exp_name": "pythia_SFT_ga4",
-  "batch_size": 64,
-  "eval_batch_size": 32,
-  "debug": false,
-  "fsdp_port": 12355,
-  "datasets": [
-    "hh"
+  "architectures": [
+    "GPTNeoXForCausalLM"
   ],
-  "wandb": {
-    "enabled": true,
-    "entity": "pythia_dpo",
-    "project": "Pythia_LOM"
-  },
-  "local_dirs": [
-    "/scr-ssd",
-    "/scr",
-    ".cache"
-  ],
-  "sample_during_eval": false,
-  "n_eval_model_samples": 16,
-  "do_first_eval": true,
-  "local_run_dir": ".cache/laura/pythia_SFT_ga4_2023-07-16_16-50-13_244945",
-  "lr": 0.000001,
-  "gradient_accumulation_steps": 4,
-  "max_grad_norm": 10,
-  "max_length": 512,
-  "max_prompt_length": 256,
-  "n_epochs": 1,
-  "n_examples": null,
-  "n_eval_examples": 256,
-  "trainer": "FSDPTrainer",
-  "optimizer": "RMSprop",
-  "warmup_steps": 150,
-  "activation_checkpointing": false,
-  "eval_every": 19968,
-  "minimum_log_interval_secs": 1,
-  "model": {
-    "name_or_path": "EleutherAI/pythia-70m",
-    "tokenizer_name_or_path": null,
-    "archive": null,
-    "block_name": "GPTNeoXLayer",
-    "policy_dtype": "float32",
-    "fsdp_policy_mp": "bfloat16",
-    "reference_dtype": "float16"
-  },
-  "loss": {
-    "name": "sft"
-  }
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_size": 512,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 6,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.24.0",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50304
 }