neurocache weights trained on long documents from red-pajama

neurocache_adapter/adapter_config.json

@@ -0,0 +1,36 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": null,
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": false,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": [
+    18,
+    19,
+    20,
+    21,
+    22,
+    23
+  ],
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "fc2",
+    "fc1"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

neurocache_config.json

@@ -10,11 +10,11 @@
   "auto_mapping": null,
   "base_model_name_or_path": "microsoft/phi-1_5",
   "cache_dtype": "bfloat16",
-  "cache_size": 16384,
+  "cache_size": 524288,
   "cache_type": "FIFO",
   "compression_factor": 4,
   "context_size": 2,
-  "global_cache": false,
+  "global_cache": true,
   "inference_mode": true,
   "neighborhood_size": 2,
   "neurocache_type": "ONDEVICE",
@@ -24,4 +24,4 @@
   "similarity_fn": "l2",
   "task_type": "CAUSAL_LM",
   "topk": 16
-}
+}

neurocache_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c78724bf9a66ff01d7cf011b3054533103f6d65886e76bda7d329bbf3ffeaa2b
+oid sha256:4c08bbc22b8e9b4e668cf7c5de2273afcd65c1bc54cd5173d15eb2cf82024316
 size 271603948