replaced ar+nar-tts+stt-llama-8 as the base ar+nar-llama-8 since it pretty much surpasses it now

models/ckpt/ar+nar-llama-8/fp32.pth

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:aca5e72710bcff66244d823b8bb887ebb48621bd74fbe8aacdac7e21cb2e7823
-size 455756877

models/ckpt/ar+nar-llama-8/fp32.sft

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2bbafd8afb5403c206c28f51ea3e872769dab8de99b5f441825ff31c893b0911
-size 455745602
+oid sha256:0920e7eef0884631f00513b700924538db0853d662530e4bcf7ac1d8666430b6
+size 456274402

models/ckpt/ar+nar-tts+stt-llama-8/fp32.sft

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0920e7eef0884631f00513b700924538db0853d662530e4bcf7ac1d8666430b6
-size 456274402

models/config.llama.tts+stt.yaml

@@ -1,130 +0,0 @@
-sample_rate: 24_000
-audio_backend: "vocos"
-
-models:
-- name: "ar+nar-tts+stt"
-  size: "full"
-  resp_levels: 8
-  prom_levels: 8
-  tasks: 9
-  langs: 4
-  tones: 1
-  arch_type: llama
-  training: False
-  version: 5
-  attention: auto
-  dropout: 0.1
-  #loss_factors:
-  #  text: 0.01
-  #  prom: 0.5
-  #  resp: 1.0
-  capabilities: ["ar", "nar"]
-  experimental:
-    p_rvq_levels: "auto"
-    audio_embedding_sums: True
-    unified_position_ids: False
-    split_classifiers: True
-    # 
-    causal_size: 1
-    interleave: False
-    rvq_level_range: []
-    tie_classifier_to_embedding: False
-
-#loras:
-#- name : "lora"
-#  rank: 128
-#  alpha: 128
-#  training: True
-#  rvq_levels: []
-
-hyperparameters:
-  batch_size: 32
-  gradient_accumulation_steps: 8
-  gradient_clipping: 1.0
-  warmup_steps: 10
-
-  optimizer: Prodigy
-  learning_rate: 1.0
-  torch_optimizer: True
-  
-  scheduler: "" # ScheduleFree
-  torch_scheduler: True
-
-evaluation:
-  batch_size: 4
-  frequency: 250
-  size: 4
-  
-  steps: 500
-  ar_temperature: 1.0
-  nar_temperature: 0.0
-
-trainer:
-  iterations: 1_000_000  
-  save_frequency: 250
-  keep_last_checkpoints: 4
-
-  resize_modules: True
-  
-  check_for_oom: False
-  gradient_checkpointing: True
-
-  weight_dtype: bfloat16
-  amp: True
-
-  backend: deepspeed
-  deepspeed:
-    inferencing: False
-    amp: False
-
-  load_webui: False
-
-inference:
-  backend: local
-  normalize: False
-
-  weight_dtype: bfloat16
-  amp: True
-
-optimizations:
-  injects: False
-  replace: True
-
-  linear: False
-  embedding: False
-  optimizers: True
-
-  bitsandbytes: False
-  dadaptation: False
-  bitnet: False
-  fp8: False
-
-dataset:
-  use_hdf5: True
-  hdf5_flag: r
-  
-  use_metadata: True
-  validate: True
-
-  workers: 1
-  cache: True
-
-  duration_range: [3.0, 12.0]
-
-  random_utterance: 1.0
-  max_prompts: 1
-  prompt_duration_range: [3.0, 3.0]
-  
-  max_resps: 1
-  p_resp_append: 0.25
-
-  sample_type: path # path # speaker
-  sample_order: duration
-  sample_max_duration_batch: 300
-  sample_shuffle: False
-
-  tasks_list: [ "tts", "stt" ]
-
-  training: []
-  validation: []
-  noise: []

models/config.llama.yaml

@@ -1,145 +1,130 @@
-sample_rate: 24_000
-audio_backend: "vocos"
-
-models:
-- name: "ar+nar"
-  size: "full"
-  resp_levels: 8
-  prom_levels: 8
-  tasks: 8
-  langs: 2
-  tones: 1
-  arch_type: llama
-  training: False
-  version: 5
-  attention: auto
-  dropout: 0.1
-  loss_factors:
-    text: 0.01
-    prom: 0.5
-    resp: 1.0
-  capabilities: ["ar", "nar"]
-  experimental:
-    audio_embedding_sums: False
-    unified_position_ids: False
-    split_classifiers: True
-
-hyperparameters:
-  autotune: False
-  autotune_params:
-    start_profile_step: 1
-    end_profile_step: 50
-    num_tuning_micro_batch_sizes: 8
-
-  batch_size: 16
-  gradient_accumulation_steps: 8
-  gradient_clipping: 1.0
-  warmup_steps: 250
-
-  optimizer: Prodigy
-  learning_rate: 1.0
-  torch_optimizer: True
-  
-  scheduler: "" # ScheduleFree
-  torch_scheduler: True
-
-evaluation:
-  batch_size: 16
-  frequency: 1000
-  size: 16
-  
-  steps: 500
-  ar_temperature: 0.95
-  nar_temperature: 0.25
-  load_disabled_engines: True
-
-trainer:
-  #no_logger: True
-  ddp: False
-  check_for_oom: False
-  iterations: 1_000_000
-  
-  save_tag: step
-  save_on_oom: True
-  save_on_quit: True
-  save_frequency: 500
-  export_on_save: True
-
-  keep_last_checkpoints: 8
-
-  aggressive_optimizations: False
-  load_disabled_engines: False
-  gradient_checkpointing: True
-
-  #load_state_dict: True
-  strict_loading: False
-  #load_tag: "9500"
-  #load_states: False
-  #restart_step_count: True
-  
-  gc_mode: None # "global_step"
-
-  weight_dtype: bfloat16
-  amp: True
-
-  backend: deepspeed
-  deepspeed:
-    inferencing: True
-    zero_optimization_level: 0
-    use_compression_training: False
-
-    amp: False
-
-  load_webui: False
-
-inference:
-  backend: deepspeed
-  audio_backend: "vocos"
-  normalize: False
-
-  weight_dtype: bfloat16
-  amp: True
-
-optimizations:
-  injects: False
-  replace: True
-
-  linear: False
-  embedding: False
-  optimizers: True
-
-  bitsandbytes: False
-  dadaptation: False
-  bitnet: False
-  fp8: False
-
-dataset:
-  speaker_name_getter: "lambda p: f'{p.parts[-3]}_{p.parts[-2]}'"
-  speaker_group_getter: "lambda p: f'{p.parts[-3]}'"
-  speaker_languages:
-    ja: []
-
-  use_hdf5: True
-  use_metadata: True
-  hdf5_flag: r
-  validate: True
-
-  workers: 6
-  cache: True
-
-  duration_range: [24.0, 32.0]
-
-  random_utterance: 1.0
-  max_prompts: 1
-  prompt_duration_range: [3.0, 9.0]
-  
-  max_resps: 1
-  p_resp_append: 0.25
-
-  sample_type: path # path # speaker
-
-  tasks_list: [ "tts" ] # , [ "tts", "tts-c", "ns", "sr", "tse", "cse", "nse", "tts"]
-
-  training: []
-  validation: []
-  noise: []
+sample_rate: 24_000
+audio_backend: "vocos"
+
+models:
+- name: "ar+nar-tts+stt"
+  size: "full"
+  resp_levels: 8
+  prom_levels: 8
+  tasks: 9
+  langs: 4
+  tones: 1
+  arch_type: llama
+  training: False
+  version: 5
+  attention: auto
+  dropout: 0.1
+  #loss_factors:
+  #  text: 0.01
+  #  prom: 0.5
+  #  resp: 1.0
+  capabilities: ["ar", "nar"]
+  experimental:
+    p_rvq_levels: "auto"
+    audio_embedding_sums: True
+    unified_position_ids: False
+    split_classifiers: True
+    # 
+    causal_size: 1
+    interleave: False
+    rvq_level_range: []
+    tie_classifier_to_embedding: False
+
+#loras:
+#- name : "lora"
+#  rank: 128
+#  alpha: 128
+#  training: True
+#  rvq_levels: []
+
+hyperparameters:
+  batch_size: 32
+  gradient_accumulation_steps: 8
+  gradient_clipping: 1.0
+  warmup_steps: 10
+
+  optimizer: Prodigy
+  learning_rate: 1.0
+  torch_optimizer: True
+  
+  scheduler: "" # ScheduleFree
+  torch_scheduler: True
+
+evaluation:
+  batch_size: 4
+  frequency: 250
+  size: 4
+  
+  steps: 500
+  ar_temperature: 1.0
+  nar_temperature: 0.0
+
+trainer:
+  iterations: 1_000_000  
+  save_frequency: 250
+  keep_last_checkpoints: 4
+
+  resize_modules: True
+  
+  check_for_oom: False
+  gradient_checkpointing: True
+
+  weight_dtype: bfloat16
+  amp: True
+
+  backend: deepspeed
+  deepspeed:
+    inferencing: False
+    amp: False
+
+  load_webui: False
+
+inference:
+  backend: local
+  normalize: False
+
+  weight_dtype: bfloat16
+  amp: True
+
+optimizations:
+  injects: False
+  replace: True
+
+  linear: False
+  embedding: False
+  optimizers: True
+
+  bitsandbytes: False
+  dadaptation: False
+  bitnet: False
+  fp8: False
+
+dataset:
+  use_hdf5: True
+  hdf5_flag: r
+  
+  use_metadata: True
+  validate: True
+
+  workers: 1
+  cache: True
+
+  duration_range: [3.0, 12.0]
+
+  random_utterance: 1.0
+  max_prompts: 1
+  prompt_duration_range: [3.0, 3.0]
+  
+  max_resps: 1
+  p_resp_append: 0.25
+
+  sample_type: path # path # speaker
+  sample_order: duration
+  sample_max_duration_batch: 300
+  sample_shuffle: False
+
+  tasks_list: [ "tts", "stt" ]
+
+  training: []
+  validation: []
+  noise: []