Upload folder using huggingface_hub

README.md

@@ -30,8 +30,8 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
 
 # Instantiate model and tokenizer
-tokenizer = AutoTokenizer.from_pretrained("Zyphra/Zamba2-7B-Instruct")
-model = AutoModelForCausalLM.from_pretrained("Zyphra/Zamba2-7B-Instruct", device_map="cuda", torch_dtype=torch.bfloat16)
+tokenizer = AutoTokenizer.from_pretrained("Zyphra/Zamba2-7B-instruct")
+model = AutoModelForCausalLM.from_pretrained("Zyphra/Zamba2-7B-instruct", device_map="cuda", torch_dtype=torch.bfloat16)
 
 # Format the input as a chat template
 user_turn_1 = "In one season a flower blooms three times. In one year, there is one blooming season. How many times do two flowers bloom in two years? Please include your logic."

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4e5e9d53a3156268f039fcb18b65b8ce4b88e7348be51f163c3a7c7e0cb7a1a3
-size 4917529496
+oid sha256:8fb157a960ba1569c04c4d2e141d75b49fa081dd3639534cc2ef1c1d644e0272
+size 4917529456

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:908e9a85c51bd1de0fa3883cba1d5225f2caf579bc5c935e0759a5a366e32341
-size 4968641256
+oid sha256:877a5e24da880f00d15f1a2c190df9883f0fae33139b548bae62012c7cc73f3c
+size 4968641192

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ad3678f9b8e71d1ad181fe092fa1c1ed91d83319a0342de1ff885399c4e51c3
-size 4934769840
+oid sha256:7d40e600ea6b9b3553b5af4cdde461fca9d0fc4b99d0f08d66488c642215733e
+size 4934769784

model.safetensors.index.json

@@ -93,7 +93,7 @@
  "model.mamba_layers.0.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.0.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.0.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.0.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.0.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.0.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.0.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -102,7 +102,7 @@
  "model.mamba_layers.1.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.1.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.1.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.1.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.1.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.1.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.1.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.10.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -111,7 +111,7 @@
  "model.mamba_layers.10.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.10.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.10.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.10.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.10.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.10.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.10.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.11.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -120,7 +120,7 @@
  "model.mamba_layers.11.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.11.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.11.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.11.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.11.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.11.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.11.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.12.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -129,7 +129,7 @@
  "model.mamba_layers.12.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.12.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.12.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.12.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.12.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.12.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.12.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.13.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -138,7 +138,7 @@
  "model.mamba_layers.13.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.13.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.13.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.13.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.13.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.13.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.13.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.14.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -147,7 +147,7 @@
  "model.mamba_layers.14.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.14.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.14.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.14.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.14.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.14.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.14.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.15.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -156,7 +156,7 @@
  "model.mamba_layers.15.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.15.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.15.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.15.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.15.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.15.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.15.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.16.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -165,7 +165,7 @@
  "model.mamba_layers.16.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.16.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.16.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.16.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.16.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.16.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.16.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.17.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -174,7 +174,7 @@
  "model.mamba_layers.17.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.17.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.17.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.17.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.17.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.17.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.17.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.18.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -183,7 +183,7 @@
  "model.mamba_layers.18.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.18.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.18.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.18.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.18.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.18.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.18.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.19.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -192,7 +192,7 @@
  "model.mamba_layers.19.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.19.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.19.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.19.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.19.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.19.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.19.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -201,7 +201,7 @@
  "model.mamba_layers.2.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.2.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.2.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.2.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.2.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.2.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.2.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -210,7 +210,7 @@
  "model.mamba_layers.20.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.20.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.20.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.20.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.20.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.20.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.20.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.21.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -219,7 +219,7 @@
  "model.mamba_layers.21.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.21.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.21.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.21.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.21.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.21.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.21.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.22.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -228,7 +228,7 @@
  "model.mamba_layers.22.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.22.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.22.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.22.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.22.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.22.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.22.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.23.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -237,7 +237,7 @@
  "model.mamba_layers.23.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.23.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.23.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.23.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.23.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.23.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.23.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.24.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -246,7 +246,7 @@
  "model.mamba_layers.24.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.24.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.24.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.24.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.24.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.24.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.24.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.25.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -255,7 +255,7 @@
  "model.mamba_layers.25.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.25.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.25.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.25.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.25.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.25.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.25.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.26.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -264,7 +264,7 @@
  "model.mamba_layers.26.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.26.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.26.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.26.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.26.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.26.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.26.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.27.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -273,7 +273,7 @@
  "model.mamba_layers.27.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.27.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.27.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.27.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.27.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.27.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.27.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.28.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -282,7 +282,7 @@
  "model.mamba_layers.28.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.28.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.28.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.28.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.28.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.28.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.28.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.29.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -291,7 +291,7 @@
  "model.mamba_layers.29.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.29.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.29.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.29.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.29.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.29.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.29.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -300,7 +300,7 @@
  "model.mamba_layers.3.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.3.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.3.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.3.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.3.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.3.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.3.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.30.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -309,7 +309,7 @@
  "model.mamba_layers.30.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.30.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.30.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.30.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.30.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.30.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.30.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.31.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -318,7 +318,7 @@
  "model.mamba_layers.31.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.31.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.31.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.31.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.31.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.31.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.31.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.32.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -327,7 +327,7 @@
  "model.mamba_layers.32.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.32.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.32.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.32.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.32.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.32.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.32.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.33.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -336,7 +336,7 @@
  "model.mamba_layers.33.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.33.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.33.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.33.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.33.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.33.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.33.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.34.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -345,7 +345,7 @@
  "model.mamba_layers.34.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.34.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.34.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.34.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.34.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.34.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.34.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.35.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -354,7 +354,7 @@
  "model.mamba_layers.35.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.35.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.35.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.35.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.35.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.35.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.35.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.36.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -363,7 +363,7 @@
  "model.mamba_layers.36.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.36.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.36.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.36.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.36.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.36.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.36.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.37.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -372,7 +372,7 @@
  "model.mamba_layers.37.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.37.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.37.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.37.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.37.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.37.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.37.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.38.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -381,7 +381,7 @@
  "model.mamba_layers.38.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.38.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.38.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.38.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.38.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.38.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.38.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.39.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -390,7 +390,7 @@
  "model.mamba_layers.39.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.39.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.39.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.39.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.39.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.39.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.39.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -399,7 +399,7 @@
  "model.mamba_layers.4.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.4.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.4.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.4.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.4.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.4.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.4.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.40.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -408,7 +408,7 @@
  "model.mamba_layers.40.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.40.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.40.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.40.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.40.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.40.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.40.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.41.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -417,7 +417,7 @@
  "model.mamba_layers.41.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.41.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.41.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.41.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.41.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.41.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.41.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.42.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -426,7 +426,7 @@
  "model.mamba_layers.42.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.42.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.42.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.42.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.42.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.42.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.42.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.43.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -435,7 +435,7 @@
  "model.mamba_layers.43.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.43.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.43.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.43.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.43.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.43.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.43.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.44.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -444,7 +444,7 @@
  "model.mamba_layers.44.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.44.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.44.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.44.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.44.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.44.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.44.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.45.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -453,7 +453,7 @@
  "model.mamba_layers.45.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.45.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.45.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.45.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.45.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.45.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.45.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.46.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -462,7 +462,7 @@
  "model.mamba_layers.46.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.46.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.46.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.46.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.46.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.46.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.46.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.47.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -471,7 +471,7 @@
  "model.mamba_layers.47.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.47.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.47.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.47.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.47.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.47.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.47.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.48.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -480,7 +480,7 @@
  "model.mamba_layers.48.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.48.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.48.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.48.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.48.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.48.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.48.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.49.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -489,7 +489,7 @@
  "model.mamba_layers.49.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.49.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.49.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.49.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.49.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.49.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.49.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -498,7 +498,7 @@
  "model.mamba_layers.5.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.5.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.5.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.5.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.5.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.5.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.5.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.50.input_layernorm.weight": "model-00002-of-00004.safetensors",
@@ -507,7 +507,7 @@
  "model.mamba_layers.50.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.50.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.50.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.50.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.50.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.50.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.50.mamba.out_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.51.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -516,7 +516,7 @@
  "model.mamba_layers.51.mamba.conv1d.bias": "model-00002-of-00004.safetensors",
  "model.mamba_layers.51.mamba.conv1d.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.51.mamba.dt_bias": "model-00002-of-00004.safetensors",
- "model.mamba_layers.51.mamba.in_proj.0.weight": "model-00002-of-00004.safetensors",
+ "model.mamba_layers.51.mamba.in_proj.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.51.mamba.norm.weight": "model-00002-of-00004.safetensors",
  "model.mamba_layers.51.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.52.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -525,7 +525,7 @@
  "model.mamba_layers.52.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.52.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.52.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.52.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.52.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.52.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.52.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.53.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -534,7 +534,7 @@
  "model.mamba_layers.53.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.53.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.53.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.53.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.53.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.53.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.53.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.54.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -543,7 +543,7 @@
  "model.mamba_layers.54.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.54.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.54.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.54.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.54.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.54.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.54.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.55.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -552,7 +552,7 @@
  "model.mamba_layers.55.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.55.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.55.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.55.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.55.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.55.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.55.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.56.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -561,7 +561,7 @@
  "model.mamba_layers.56.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.56.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.56.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.56.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.56.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.56.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.56.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.57.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -570,7 +570,7 @@
  "model.mamba_layers.57.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.57.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.57.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.57.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.57.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.57.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.57.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.58.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -579,7 +579,7 @@
  "model.mamba_layers.58.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.58.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.58.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.58.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.58.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.58.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.58.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.59.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -588,7 +588,7 @@
  "model.mamba_layers.59.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.59.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.59.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.59.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.59.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.59.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.59.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -597,7 +597,7 @@
  "model.mamba_layers.6.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.6.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.6.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.6.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.6.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.6.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.6.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.60.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -606,7 +606,7 @@
  "model.mamba_layers.60.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.60.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.60.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.60.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.60.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.60.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.60.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.61.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -615,7 +615,7 @@
  "model.mamba_layers.61.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.61.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.61.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.61.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.61.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.61.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.61.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.62.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -624,7 +624,7 @@
  "model.mamba_layers.62.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.62.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.62.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.62.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.62.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.62.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.62.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.63.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -633,7 +633,7 @@
  "model.mamba_layers.63.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.63.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.63.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.63.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.63.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.63.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.63.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.64.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -642,7 +642,7 @@
  "model.mamba_layers.64.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.64.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.64.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.64.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.64.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.64.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.64.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.65.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -651,7 +651,7 @@
  "model.mamba_layers.65.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.65.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.65.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.65.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.65.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.65.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.65.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.66.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -660,7 +660,7 @@
  "model.mamba_layers.66.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.66.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.66.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.66.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.66.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.66.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.66.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.67.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -669,7 +669,7 @@
  "model.mamba_layers.67.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.67.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.67.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.67.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.67.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.67.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.67.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.68.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -678,7 +678,7 @@
  "model.mamba_layers.68.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.68.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.68.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.68.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.68.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.68.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.68.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.69.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -687,7 +687,7 @@
  "model.mamba_layers.69.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.69.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.69.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.69.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.69.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.69.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.69.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -696,7 +696,7 @@
  "model.mamba_layers.7.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.7.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.7.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.7.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.7.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.7.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.7.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.70.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -705,7 +705,7 @@
  "model.mamba_layers.70.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.70.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.70.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.70.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.70.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.70.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.70.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.71.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -714,7 +714,7 @@
  "model.mamba_layers.71.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.71.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.71.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.71.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.71.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.71.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.71.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.72.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -723,7 +723,7 @@
  "model.mamba_layers.72.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.72.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.72.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.72.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.72.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.72.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.72.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.73.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -732,7 +732,7 @@
  "model.mamba_layers.73.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.73.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.73.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.73.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.73.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.73.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.73.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.74.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -741,7 +741,7 @@
  "model.mamba_layers.74.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.74.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.74.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.74.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.74.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.74.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.74.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.75.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -750,7 +750,7 @@
  "model.mamba_layers.75.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.75.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.75.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.75.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.75.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.75.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.75.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.76.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -759,7 +759,7 @@
  "model.mamba_layers.76.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.76.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.76.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.76.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.76.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.76.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.76.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.77.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -768,7 +768,7 @@
  "model.mamba_layers.77.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.77.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.77.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.77.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.77.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.77.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.77.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.78.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -777,7 +777,7 @@
  "model.mamba_layers.78.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.78.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.78.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.78.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.78.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.78.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.78.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.79.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -786,7 +786,7 @@
  "model.mamba_layers.79.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.79.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.79.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.79.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.79.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.79.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.79.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -795,7 +795,7 @@
  "model.mamba_layers.8.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.8.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.8.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.8.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.8.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.8.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.8.mamba.out_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.80.input_layernorm.weight": "model-00003-of-00004.safetensors",
@@ -804,7 +804,7 @@
  "model.mamba_layers.80.mamba.conv1d.bias": "model-00003-of-00004.safetensors",
  "model.mamba_layers.80.mamba.conv1d.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.80.mamba.dt_bias": "model-00003-of-00004.safetensors",
- "model.mamba_layers.80.mamba.in_proj.0.weight": "model-00003-of-00004.safetensors",
+ "model.mamba_layers.80.mamba.in_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.80.mamba.norm.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.80.mamba.out_proj.weight": "model-00003-of-00004.safetensors",
  "model.mamba_layers.9.input_layernorm.weight": "model-00001-of-00004.safetensors",
@@ -813,8 +813,8 @@
  "model.mamba_layers.9.mamba.conv1d.bias": "model-00001-of-00004.safetensors",
  "model.mamba_layers.9.mamba.conv1d.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.9.mamba.dt_bias": "model-00001-of-00004.safetensors",
- "model.mamba_layers.9.mamba.in_proj.0.weight": "model-00001-of-00004.safetensors",
+ "model.mamba_layers.9.mamba.in_proj.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.9.mamba.norm.weight": "model-00001-of-00004.safetensors",
  "model.mamba_layers.9.mamba.out_proj.weight": "model-00001-of-00004.safetensors"
  }
-}
+}