Update README.md

README.md

@@ -17,23 +17,23 @@ Trained over 73+ hours of "train" split of ylacombe/cml-tts dataset
 with 8xRTX4090, still in progress, using gradio finetuning app using following settings:
 ```
 exp_name"F5TTS_Base"
-learning_rate0.00001
-batch_size_per_gpu10000
-batch_size_type"frame"
-max_samples64
-grad_accumulation_steps1
-max_grad_norm1
-epochs100
-num_warmup_updates2000
-save_per_updates600
-last_per_steps300
-finetunetrue
-file_checkpoint_train""
-tokenizer_type"char"
-tokenizer_file""
-mixed_precision"fp16"
-logger"wandb"
-bnb_optimizerfalse
+learning_rate=0.00001
+batch_size_per_gpu=10000
+batch_size_type="frame"
+max_samples=64
+grad_accumulation_steps=1
+max_grad_norm=1
+epochs=300
+num_warmup_updates=2000
+save_per_updates=600
+last_per_steps=300
+finetune=true
+file_checkpoint_train=""
+tokenizer_type="char"
+tokenizer_file=""
+mixed_precision="fp16"
+logger="wandb"
+bnb_optimizer=false
 ```
 
 # Pre processing 
@@ -46,9 +46,16 @@ I'm only talking about Italian data on cml-tts, I don't know if other languages
 
 
 # Current most trained model
-model_25200.safetensors (45 Epoch)
+model_159600.safetensors (~290 Epoch)
+
+## known problems
+- catastrophic failure (being Italian only, lost english skill). A proper multilanguage dataset should be used instead of single language.
+- not perfect pronunciation
+- numbers must be converter in letters to be pronunced in italian
+- a better dataset with more diverse voices would help improving zero-shot cloning
 
 
 ### checkpoints folder
 Contains the weight of the checkpoints at specific steps, the higher the number, the further it went into training.
-Weights in this folder can be used as starting point to continue training.
+Weights in this folder can be used as starting point to continue training.
+Ping me back if you can further finetune it to reach a better result