diff --git a/aiserver.py b/aiserver.py
index e61cc3b7..0f3a10c5 100644
--- a/aiserver.py
+++ b/aiserver.py
@@ -2491,14 +2491,15 @@ def load_model(use_gpu=True, gpu_layers=None, disk_layers=None, initial_load=Fal
                     if(os.path.isdir(vars.custmodpth)):
                         try:
                             tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
+                        except Exception as e:
+                            pass
+                        try:
+                            tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache", use_fast=False)
                         except Exception as e:
                             try:
-                                tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache", use_fast=False)
+                                tokenizer = GPT2TokenizerFast.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
                             except Exception as e:
-                                try:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
-                                except Exception as e:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                                tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
                         try:
                             model     = AutoModelForCausalLM.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache", **lowmem)
                         except Exception as e:
@@ -2508,14 +2509,15 @@ def load_model(use_gpu=True, gpu_layers=None, disk_layers=None, initial_load=Fal
                     elif(os.path.isdir("models/{}".format(vars.model.replace('/', '_')))):
                         try:
                             tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
+                        except Exception as e:
+                            pass
+                        try:
+                            tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache", use_fast=False)
                         except Exception as e:
                             try:
-                                tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache", use_fast=False)
+                                tokenizer = GPT2TokenizerFast.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
                             except Exception as e:
-                                try:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
-                                except Exception as e:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                                tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
                         try:
                             model     = AutoModelForCausalLM.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache", **lowmem)
                         except Exception as e:
@@ -2538,14 +2540,15 @@ def load_model(use_gpu=True, gpu_layers=None, disk_layers=None, initial_load=Fal
 
                         try:
                             tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
+                        except Exception as e:
+                            pass
+                        try:
+                            tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache", use_fast=False)
                         except Exception as e:
                             try:
-                                tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache", use_fast=False)
+                                tokenizer = GPT2TokenizerFast.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
                             except Exception as e:
-                                try:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
-                                except Exception as e:
-                                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                                tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
                         try:
                             model     = AutoModelForCausalLM.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache", **lowmem)
                         except Exception as e:
diff --git a/tpu_mtj_backend.py b/tpu_mtj_backend.py
index 9c1466f7..79582867 100644
--- a/tpu_mtj_backend.py
+++ b/tpu_mtj_backend.py
@@ -1352,14 +1352,15 @@ def load_model(path: str, driver_version="tpu_driver0.1_dev20210607", hf_checkpo
         if(os.path.isdir(vars.custmodpth)):
             try:
                 tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
+            except Exception as e:
+                pass
+            try:
+                tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache", use_fast=False)
             except Exception as e:
                 try:
-                    tokenizer = AutoTokenizer.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache", use_fast=False)
+                    tokenizer = GPT2TokenizerFast.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
                 except Exception as e:
-                    try:
-                        tokenizer = GPT2TokenizerFast.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
-                    except Exception as e:
-                        tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
             try:
                 model     = AutoModelForCausalLM.from_pretrained(vars.custmodpth, revision=vars.revision, cache_dir="cache")
             except Exception as e:
@@ -1367,14 +1368,15 @@ def load_model(path: str, driver_version="tpu_driver0.1_dev20210607", hf_checkpo
         elif(os.path.isdir("models/{}".format(vars.model.replace('/', '_')))):
             try:
                 tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
+            except Exception as e:
+                pass
+            try:
+                tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache", use_fast=False)
             except Exception as e:
                 try:
-                    tokenizer = AutoTokenizer.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache", use_fast=False)
+                    tokenizer = GPT2TokenizerFast.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
                 except Exception as e:
-                    try:
-                        tokenizer = GPT2TokenizerFast.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
-                    except Exception as e:
-                        tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
             try:
                 model     = AutoModelForCausalLM.from_pretrained("models/{}".format(vars.model.replace('/', '_')), revision=vars.revision, cache_dir="cache")
             except Exception as e:
@@ -1382,14 +1384,15 @@ def load_model(path: str, driver_version="tpu_driver0.1_dev20210607", hf_checkpo
         else:
             try:
                 tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
+            except Exception as e:
+                pass
+            try:
+                tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache", use_fast=False)
             except Exception as e:
                 try:
-                    tokenizer = AutoTokenizer.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache", use_fast=False)
+                    tokenizer = GPT2TokenizerFast.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
                 except Exception as e:
-                    try:
-                        tokenizer = GPT2TokenizerFast.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
-                    except Exception as e:
-                        tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
+                    tokenizer = GPT2TokenizerFast.from_pretrained("gpt2", revision=vars.revision, cache_dir="cache")
             try:
                 model     = AutoModelForCausalLM.from_pretrained(vars.model, revision=vars.revision, cache_dir="cache")
             except Exception as e: