fix: prep index for mem attn

configs/train/finetune_memory.yaml

@@ -0,0 +1,45 @@
+# model/tokenizer
+model_name: "EleutherAI/pythia-1b"
+tokenizer_name: "EleutherAI/pythia-1b"
+version: null
+gradient_checkpointing: true
+save_name: "nomic-ai/lethe"
+push_to_hub: false
+memory_attn_layer: 12
+
+# dataset
+streaming: false
+num_proc: 64
+dataset_path: "/home/paperspace/gpt4all/gpt4all/index/squad_supplemented_train" 
+max_length: 1024
+batch_size: 64
+pct_test: 0.05
+q_column: "question"
+a_column: "answers"
+context_column: "neighbor_text"
+num_memories_per_index: 500000
+num_neighbors_to_retrieve: 32
+mem_chunk_size: 64
+
+# train dynamics
+lr: 1.0e-4
+min_lr: 0 
+weight_decay: 0.0
+eval_every: 100 
+save_every: -1
+log_grads_every: 100
+log_lr_every: 10
+output_dir: "ckpts/mem_attn"
+checkpoint: null
+lora: false
+warmup_steps: 500
+num_epochs: 5 
+debug: false
+scheduler: false
+
+# logging
+wandb: false
+wandb_entity: gpt4all
+wandb_project_name: mem_attn
+seed: 42
+

gpt4all/index/prep_index_for_train.py

@@ -11,7 +11,7 @@ from gpt4all.utils.read import read_config
 from gpt4all.index.embed_texts import embed_texts
 from tqdm import tqdm
 
-CHUNK_SIZE = 1024
+CHUNK_SIZE = 2048
 
 def parse_args():
     parser = ArgumentParser()
@@ -57,10 +57,10 @@ def prep_index():
     #build training dataset
     train_dataset = load_dataset(ds_path, split=args.split)
 
-    #search the index for each query
-    neighbor_embs_column = []
+    # search the index for each query
+    neighbor_text_column = []
     neighbor_ids_column = []
-    for chunk_start in tqdm(range(0, len(ds), CHUNK_SIZE)):
+    for _, chunk_start in enumerate(tqdm(range(0, len(ds), CHUNK_SIZE))):
         chunk_end = chunk_start + CHUNK_SIZE
         chunk = ds[chunk_start:chunk_end]
         query_vectors = np.array(chunk['embedding'])
@@ -70,21 +70,21 @@ def prep_index():
 
         # build mapping between indices and embeddings
         neighbor_id_list = neighbor_objs['id']
-        emb_list = neighbor_objs['embedding']
-        idx_to_embedding = {idx.as_py(): emb_list[i] for i, idx in enumerate(neighbor_id_list)}
+        documents = neighbor_objs['text']
+        idx_to_embedding = {idx.as_py(): documents[i] for i, idx in enumerate(neighbor_id_list)}
 
-        neighbor_embs = []
+        neighbor_text = []
         for cur_neighbor_ids in neighbor_ids:
             cur_embs = [idx_to_embedding[id].as_py() for id in cur_neighbor_ids]
-            neighbor_embs.append(cur_embs)
+            neighbor_text.append(cur_embs)
 
-        neighbor_embs_column.extend(neighbor_embs)
+        neighbor_text_column.extend(neighbor_text)
         neighbor_ids_column.extend(neighbor_ids)
 
     print("adding neighbor ids")
     train_dataset = train_dataset.add_column('neighbor_ids', neighbor_ids_column)
-    print("adding neighbor embeddings")
-    train_dataset = train_dataset.add_column('neighbor_embeddings', neighbor_embs_column)
+    print("adding neighbors")
+    train_dataset = train_dataset.add_column('neighbor_text', neighbor_text_column)
 
     supplemented_dataset_path = f"{ds_path}_supplemented_{args.split}/"
     train_dataset.save_to_disk(supplemented_dataset_path)