adding guard statements

src/axolotl/cli/integrations/convert_diff_transformer.py

@@ -50,6 +50,13 @@ def test_inference(model, tokenizer, prompt="The quick brown fox"):
 
 
 def convert_diff_transformer(cfg, cli_args, config_path):
+    assert not (
+        cli_args.split_heads and cli_args.zero_init
+    ), "Both `split_heads` and `zero_init` cannot be `True`"
+    assert not (
+        cli_args.zero_init and cli_args.mirror_weights
+    ), "Both `zero_init` and `mirror_weights` cannot be `True`"
+
     debug_info = {}
 
     # Load model and tokenizer
@@ -72,21 +79,16 @@ def convert_diff_transformer(cfg, cli_args, config_path):
             model, tokenizer
         )
 
-    # Convert attention
-    LOG.info("Converting to differential attention...")
-    if cli_args.split_heads and cli_args.zero_init:
-        LOG.warning(
-            Fore.YELLOW
-            + "Warning: Using split_heads with zero_init is not recommended; "
-            + "split_heads will preclude the effects of zero_init"
-            + Fore.RESET
-        )
     try:
+        # Convert attention
+        LOG.info("Converting to differential attention...")
+
         config = LlamaDifferentialConfig(
             **model.config.__dict__,
             zero_init=cli_args.zero_init,
             sublayer_norm=cli_args.sublayer_norm,
             split_heads=cli_args.split_heads,
+            mirror_weights=cli_args.mirror_weights,
         )
         model = LlamaDifferentialForCausalLM.from_llama(model, config)
         model.to(cfg.device, dtype=cfg.torch_dtype)

src/axolotl/common/cli.py

@@ -56,6 +56,7 @@ class ConvertDiffTransformerCliArgs:
     zero_init: bool = field(default=False)
     sublayer_norm: bool = field(default=True)
     split_heads: bool = field(default=False)
+    mirror_weights: bool = field(default=False)
 
 
 def load_model_and_tokenizer(

src/axolotl/integrations/diff_transformer/modeling_diff_attn.py

@@ -36,6 +36,7 @@ class LlamaDifferentialConfig(LlamaConfig):
         split_heads: bool = False,
         sublayer_norm: bool = True,
         zero_init: bool = False,
+        mirror_weights: bool = False,
         **kwargs,
     ):
         """
@@ -45,12 +46,15 @@ class LlamaDifferentialConfig(LlamaConfig):
             split_heads: Whether to use split heads mode for attention computation.
             sublayer_norm: Whether to apply normalization to sublayers.
             zero_init: Whether to initialize new weights to zero.
+            mirror_weights: Whether to copy the positive attention component weights to
+                the negative attention component.
             **kwargs: Additional arguments passed to LlamaConfig.
         """
         super().__init__(**kwargs)
         self.split_heads = split_heads
         self.sublayer_norm = sublayer_norm
         self.zero_init = zero_init
+        self.mirror_weights = mirror_weights
         self.architectures = ["LlamaDifferentialModel"]
         self._attn_implementations = {
             "eager": "differential_eager",
@@ -250,7 +254,7 @@ class LlamaDifferentialModel(LlamaModel):
                         new_layer.self_attn.lambda_q2.zero_()
                         new_layer.self_attn.lambda_k2.zero_()
                         new_layer.self_attn.lambda_init.zero_()
-                else:
+                elif config.mirror_weights:
                     # Mirror weights for second component
                     new_layer.self_attn.q_proj.weight.data[old_q_size:].copy_(
                         old_layer.self_attn.q_proj.weight.data