Merge pull request #48 from stanfordnlp/zen/cachepath

frankaging · web-flow · commit 7a582559526d · 2024-01-12T21:17:20.000-08:00
Remove default cache directory
diff --git a/pyvene/models/blip/modelings_intervenable_blip.py b/pyvene/models/blip/modelings_intervenable_blip.py
@@ -99,7 +99,7 @@
 blip_wrapper_type_to_dimension_mapping = blip_type_to_dimension_mapping
 
 
-def create_blip(name="Salesforce/blip-vqa-base", cache_dir="../../.huggingface_cache"):
+def create_blip(name="Salesforce/blip-vqa-base", cache_dir=None):
     """Creates a GPT2 model, config, and tokenizer from the given name and revision"""
     from transformers import BlipConfig, BlipProcessor, BlipForQuestionAnswering
 
diff --git a/pyvene/models/gpt2/modelings_intervenable_gpt2.py b/pyvene/models/gpt2/modelings_intervenable_gpt2.py
@@ -63,7 +63,7 @@
 gpt2_lm_type_to_dimension_mapping = gpt2_type_to_dimension_mapping
 
 
-def create_gpt2(name="gpt2", cache_dir="../../.huggingface_cache"):
+def create_gpt2(name="gpt2", cache_dir=None):
     """Creates a GPT2 model, config, and tokenizer from the given name and revision"""
     from transformers import GPT2Model, GPT2Tokenizer, GPT2Config
 
@@ -74,7 +74,7 @@ def create_gpt2(name="gpt2", cache_dir="../../.huggingface_cache"):
     return config, tokenizer, gpt
 
 
-def create_gpt2_lm(name="gpt2", config=None, cache_dir="../../.huggingface_cache"):
+def create_gpt2_lm(name="gpt2", config=None, cache_dir=None):
     """Creates a GPT2 LM, config, and tokenizer from the given name and revision"""
     from transformers import GPT2LMHeadModel, GPT2Tokenizer, GPT2Config
 
diff --git a/pyvene/models/gpt_neo/modelings_intervenable_gpt_neo.py b/pyvene/models/gpt_neo/modelings_intervenable_gpt_neo.py
@@ -64,7 +64,7 @@
 
 
 def create_gpt_neo(
-    name="roneneldan/TinyStories-33M", cache_dir="../../.huggingface_cache"
+    name="roneneldan/TinyStories-33M", cache_dir=None
 ):
     """Creates a GPT2 model, config, and tokenizer from the given name and revision"""
     from transformers import GPTNeoForCausalLM, GPT2Tokenizer, GPTNeoConfig
diff --git a/pyvene/models/gpt_neox/modelings_intervenable_gpt_neox.py b/pyvene/models/gpt_neox/modelings_intervenable_gpt_neox.py
@@ -63,7 +63,7 @@
 gpt_neox_lm_type_to_dimension_mapping = gpt_neox_type_to_dimension_mapping
 
 
-def create_gpt_neox(name="EleutherAI/pythia-70m", cache_dir="../../.huggingface_cache"):
+def create_gpt_neox(name="EleutherAI/pythia-70m", cache_dir=None):
     """Creates a GPT2 model, config, and tokenizer from the given name and revision"""
     from transformers import GPTNeoXForCausalLM, AutoTokenizer, GPTNeoXConfig
 
diff --git a/pyvene/models/gru/modelings_intervenable_gru.py b/pyvene/models/gru/modelings_intervenable_gru.py
@@ -69,7 +69,7 @@
 gru_lm_type_to_dimension_mapping = gru_type_to_dimension_mapping
 
 
-def create_gru(config, tokenizer_name=None, cache_dir="../../.huggingface_cache"):
+def create_gru(config, tokenizer_name=None, cache_dir=None):
     """Creates a GRU model, config, and tokenizer from the given name and revision"""
     from transformers import AutoTokenizer
     from models.gru.modelings_gru import GRUModel
@@ -82,7 +82,7 @@ def create_gru(config, tokenizer_name=None, cache_dir="../../.huggingface_cache"
     return config, tokenizer, mlp
 
 
-def create_gru_lm(config, tokenizer_name=None, cache_dir="../../.huggingface_cache"):
+def create_gru_lm(config, tokenizer_name=None, cache_dir=None):
     """Creates a GRU model, config, and tokenizer from the given name and revision"""
     from transformers import AutoTokenizer
     from models.gru.modelings_gru import GRULMHeadModel
@@ -96,7 +96,7 @@ def create_gru_lm(config, tokenizer_name=None, cache_dir="../../.huggingface_cac
 
 
 def create_gru_classifier(
-    config, tokenizer_name=None, cache_dir="../../.huggingface_cache"
+    config, tokenizer_name=None, cache_dir=None
 ):
     """Creates a GRU model, config, and tokenizer from the given name and revision"""
     from transformers import AutoTokenizer
diff --git a/pyvene/models/llama/modelings_intervenable_llama.py b/pyvene/models/llama/modelings_intervenable_llama.py
@@ -70,7 +70,7 @@ def split_heads(tensor, num_heads, attn_head_size):
 
 
 def create_llama(
-    name="sharpbai/alpaca-7b-merged", cache_dir="../../.huggingface_cache"
+    name="sharpbai/alpaca-7b-merged", cache_dir=None
 ):
     """Creates a LLaMA Causal LM model, config, and tokenizer from the given name and revision"""
     from transformers import LlamaForCausalLM, LlamaTokenizer, LlamaConfig
diff --git a/pyvene/models/mlp/modelings_intervenable_mlp.py b/pyvene/models/mlp/modelings_intervenable_mlp.py
@@ -36,7 +36,7 @@
 
 
 def create_mlp_classifier(
-    config, tokenizer_name=None, cache_dir="../../.huggingface_cache"
+    config, tokenizer_name=None, cache_dir=None
 ):
     """Creates a MLP model, config, and tokenizer from the given name and revision"""
     from transformers import AutoTokenizer
diff --git a/setup.py b/setup.py
@@ -10,7 +10,7 @@
 
 setup(
     name="pyvene",
-    version="0.0.5",
+    version="0.0.6",
     description="Use Activation Intervention to Interpret Causal Mechanism of Model",
     long_description=long_description,
     long_description_content_type='text/markdown',
diff --git a/tutorials/advanced_tutorials/IOI_with_DAS.ipynb b/tutorials/advanced_tutorials/IOI_with_DAS.ipynb
diff --git a/tutorials/advanced_tutorials/Intervened_Model_Generation.ipynb b/tutorials/advanced_tutorials/Intervened_Model_Generation.ipynb
@@ -105,7 +105,7 @@
     }
    ],
    "source": [
-    "config, tokenizer, tinystory = create_gpt_neo(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, tinystory = create_gpt_neo()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Add_Activations_to_Streams.ipynb b/tutorials/basic_tutorials/Add_Activations_to_Streams.ipynb
diff --git a/tutorials/basic_tutorials/Add_New_Model_Type.ipynb b/tutorials/basic_tutorials/Add_New_Model_Type.ipynb
diff --git a/tutorials/basic_tutorials/Basic_Intervention.ipynb b/tutorials/basic_tutorials/Basic_Intervention.ipynb
diff --git a/tutorials/basic_tutorials/Complex_Intervention.ipynb b/tutorials/basic_tutorials/Complex_Intervention.ipynb
@@ -100,7 +100,7 @@
     "    scale_y_log10,\n",
     ")\n",
     "\n",
-    "config, tokenizer, gpt = create_gpt2(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, gpt = create_gpt2()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Debug_Helper.ipynb b/tutorials/basic_tutorials/Debug_Helper.ipynb
@@ -22,7 +22,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "id": "1b5fd935",
    "metadata": {},
    "outputs": [],
@@ -71,7 +71,7 @@
     "from pyvene.models.configuration_intervenable_model import (\n",
     "    IntervenableConfig, IntervenableRepresentationConfig, VanillaIntervention)\n",
     "\n",
-    "config, tokenizer, gpt = create_gpt2(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, gpt = create_gpt2()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Interchange_Intervention_Training.ipynb b/tutorials/basic_tutorials/Interchange_Intervention_Training.ipynb
@@ -92,7 +92,7 @@
     "    IntervenableConfig, IntervenableRepresentationConfig, VanillaIntervention\n",
     ")\n",
     "\n",
-    "config, tokenizer, gpt = create_gpt2(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, gpt = create_gpt2()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Load_Save_and_Share_Interventions.ipynb b/tutorials/basic_tutorials/Load_Save_and_Share_Interventions.ipynb
@@ -102,7 +102,7 @@
     "    scale_y_log10,\n",
     ")\n",
     "\n",
-    "config, tokenizer, gpt = create_gpt2(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, gpt = create_gpt2()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Nested_Intervention.ipynb b/tutorials/basic_tutorials/Nested_Intervention.ipynb
diff --git a/tutorials/basic_tutorials/Probing.ipynb b/tutorials/basic_tutorials/Probing.ipynb
@@ -94,7 +94,7 @@
     "    CollectIntervention,\n",
     ")\n",
     "\n",
-    "config, tokenizer, gpt = create_gpt2(cache_dir=\"../../../.huggingface_cache\")"
+    "config, tokenizer, gpt = create_gpt2()"
    ]
   },
   {
diff --git a/tutorials/basic_tutorials/Subspace_Partition_with_Intervention.ipynb b/tutorials/basic_tutorials/Subspace_Partition_with_Intervention.ipynb

Original file line number	Diff line number	Diff line change
`@@ -105,7 +105,7 @@`
`105`	`105`	`}`
`106`	`106`	`],`
`107`	`107`	`"source": [`
`108`		`- "config, tokenizer, tinystory = create_gpt_neo(cache_dir=\"../../../.huggingface_cache\")"`
	`108`	`+ "config, tokenizer, tinystory = create_gpt_neo()"`
`109`	`109`	`]`
`110`	`110`	`},`
`111`	`111`	`{`