Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,58 @@
+## DFloat11 Compressed Model: `google/gemma-3-27b-it`
+
+This is a **losslessly compressed** version of [`google/gemma-3-27b-it`](https://huggingface.co/google/gemma-3-27b-it) using our custom **DFloat11** format. The outputs of this compressed model are **bit-for-bit identical** to the original BFloat16 model, while reducing GPU memory consumption by approximately **30%**.
+
+### 🔍 How It Works
+
+DFloat11 compresses model weights using **Huffman coding** of BFloat16 exponent bits, combined with **hardware-aware algorithmic designs** that enable efficient on-the-fly decompression directly on the GPU. During inference, the weights remain compressed in GPU memory and are **decompressed just before matrix multiplications**, then **immediately discarded after use** to minimize memory footprint.
+
+Key benefits:
+
+* **No CPU decompression or host-device data transfer** -- all operations are handled entirely on the GPU.
+* **Decompression overhead is constant** per forward pass and **independent of batch size**, making DFloat11 increasingly efficient at larger batch sizes.
+* DFloat11 is **much faster than CPU-offloading approaches**, enabling practical deployment in memory-constrained environments.
+* At **batch size = 1**, inference is approximately **2× slower** than the original BF16 model, but the performance gap **narrows significantly** with larger batches.
+* The compression is **fully lossless**, guaranteeing that the model’s outputs are **bit-for-bit identical** to those of the original model.
+
+### 🔧 How to Use
+
+1. Install the DFloat11 pip package *(installs the CUDA kernel automatically; requires a CUDA-compatible GPU and PyTorch installed)*:
+
+    ```bash
+    pip install dfloat11[cuda12]
+    # or if you have CUDA version 11:
+    # pip install dfloat11[cuda11]
+    ```
+
+2. To use the DFloat11 model, run the following example code in Python:
+
+    ```python
+    import torch
+    from dfloat11 import DFloat11Model
+    from transformers import AutoTokenizer
+
+    model_id = "DFloat11/gemma-3-27b-it-DF11"
+
+    model = DFloat11Model.from_pretrained(model_id, device_map="auto")
+
+    tokenizer = AutoTokenizer.from_pretrained(model_id)
+    tokenizer.pad_token = tokenizer.eos_token
+
+    prompt = "Question: What is a binary tree and its applications? Answer:"
+    inputs = tokenizer(prompt, return_tensors="pt", padding=True).to(model.device)
+
+    with torch.no_grad():
+        output = model.generate(
+            **inputs,
+            max_new_tokens=256,
+            do_sample=True,
+        )
+
+    print(tokenizer.batch_decode(output, skip_special_tokens=True))
+    ```
+
+### 📄 Learn More
+
+* **Paper**: [70% Size, 100% Accuracy: Lossless LLM Compression for Efficient GPU Inference via Dynamic-Length Float](https://arxiv.org/abs/2504.11651)
+* **GitHub**: [https://github.com/LeanModels/DFloat11](https://github.com/LeanModels/DFloat11)
+* **HuggingFace**: [https://huggingface.co/DFloat11](https://huggingface.co/DFloat11)

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

config.json

@@ -0,0 +1,80 @@
+{
+  "architectures": [
+    "Gemma3ForConditionalGeneration"
+  ],
+  "boi_token_index": 255999,
+  "dfloat11_config": {
+    "bytes_per_thread": 8,
+    "pattern_dict": {
+      "language_model.model.layers.\\d+": [
+        "self_attn.q_proj",
+        "self_attn.k_proj",
+        "self_attn.v_proj",
+        "self_attn.o_proj",
+        "mlp.gate_proj",
+        "mlp.up_proj",
+        "mlp.down_proj"
+      ]
+    },
+    "threads_per_block": [
+      512
+    ],
+    "version": "0.2.0"
+  },
+  "eoi_token_index": 256000,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "image_token_index": 262144,
+  "initializer_range": 0.02,
+  "mm_tokens_per_image": 256,
+  "model_type": "gemma3",
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attn_logit_softcapping": null,
+    "cache_implementation": "hybrid",
+    "final_logit_softcapping": null,
+    "head_dim": 128,
+    "hidden_activation": "gelu_pytorch_tanh",
+    "hidden_size": 5376,
+    "initializer_range": 0.02,
+    "intermediate_size": 21504,
+    "max_position_embeddings": 131072,
+    "model_type": "gemma3_text",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 62,
+    "num_key_value_heads": 16,
+    "query_pre_attn_scalar": 168,
+    "rms_norm_eps": 1e-06,
+    "rope_local_base_freq": 10000.0,
+    "rope_scaling": {
+      "factor": 8.0,
+      "rope_type": "linear"
+    },
+    "rope_theta": 1000000.0,
+    "sliding_window": 1024,
+    "sliding_window_pattern": 6,
+    "torch_dtype": "bfloat16",
+    "use_cache": true,
+    "vocab_size": 262208
+  },
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "vision_config": {
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 896,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 27,
+    "patch_size": 14,
+    "torch_dtype": "bfloat16",
+    "vision_use_head": false
+  }
+}

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 2,
+  "cache_implementation": "hybrid",
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "pad_token_id": 0,
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "4.51.3"
+}

language_model_model_layers_0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9d4853e8e02c4b5a75b4bee75f97b06b0132382782d4d611ea63593a753f83e
+size 572348877

language_model_model_layers_1.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23f7942c8d59b1c6a1182f80bf8d91784f944af43d690e5f30a1af0028a39f94
+size 565639036

language_model_model_layers_10.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d12575acea78c9d8aa2247c9489fc36578a4e8632375296dc3e4a99a047fc35
+size 564819996

language_model_model_layers_11.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c3f4b7797a22dc2444c935d670dc6b0e1c092e0f7762c13dcf5440c7c059520
+size 562694031

language_model_model_layers_12.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3896f5aa4fe967d3267c6096f68d6f7e0d4ad5d8d2c48339be5f3c7e594c4549
+size 569048924

language_model_model_layers_13.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db4c6ae6b11346fde0a296133af73eda25da5f9d0a19be54553817b656707b13
+size 565794148

language_model_model_layers_14.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7613ad4895df9162e3a72b61d8167651006457d921c6eb0f9e8ef46e5af131aa
+size 566904990

language_model_model_layers_15.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b63f4b7dfdb146623beb8b5419eeee98d324c65f41b9f2ef84f864861ec551c
+size 569315044

language_model_model_layers_16.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e71ccc4991e60256dc354d98f8d58daef90b7e66536391ec70fa68085c1d6265
+size 569269157

language_model_model_layers_17.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be5c4e0c0fdcd898529e8e522496e6331eb0645bb9e0fe4c0dc95514bc129606
+size 565383734

language_model_model_layers_18.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a32784d9d238ee9c5fab34f3b2be84b6019314bb85922e4157e424a638ce097d
+size 568367425

language_model_model_layers_19.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03077a73e878a52f90d48e687e2eeae97341d87e598983ed868f4f09e9884ab2
+size 569713430

language_model_model_layers_2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29f47bcd2a8cc27159fbe901765ceaf434b8b42b6705bbd6b60da3baf0b98452
+size 563808968

language_model_model_layers_20.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb3e8f1e0d3f54b24c7ef679b5f0b2909f4caacdb44f199d15f46af9d1cc15ed
+size 565408161

language_model_model_layers_21.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:428bc79b6b23474a4b3f047a16dc5a5c8873f11702052e37671f3f6a4ed04a73
+size 570813724

language_model_model_layers_22.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bcd55234aa151b25894613f5703ddd8cce7b02a30f268c2a13c52d0508d90fe
+size 566611474

language_model_model_layers_23.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:785f63ef37e21fdc68216dba4ad6ae2251f1f1ca6bfe47fe3c5ae1b44334dbc4
+size 567419484

language_model_model_layers_24.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2143c26f5d71b27f10b39c05b8dda73358ec723a9d99bbdfe73f7a62b70e8b83
+size 566916611

language_model_model_layers_25.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4841eda19c651c53aebff2d5820a8ad8e6bc3506fc381c360055640533be7dca
+size 565162864

language_model_model_layers_26.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2114c7a11c4ac4c883959e71488542f2667db3c6e05f8d7c41cccf0edcb5fc93
+size 566922806

language_model_model_layers_27.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59c337142fa535555739c09bbcfe94f9594537cde2b1e13f038ea31fbb3367b8
+size 565912274

language_model_model_layers_28.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4588f999875b43b062541970f1319cbaa9ece55d1bac9fbbbc3f7fa855d93be8
+size 566539441

language_model_model_layers_29.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8071473c2619b049d54edec3089ecdb6437f58dbf66505689a85550c333501ee
+size 572270719

language_model_model_layers_3.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42d720033c8d14800e5ef0a9733caa6f1fb130fa4c465932dd023e539f0c0547
+size 567951696

language_model_model_layers_30.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0168e8ff6687cf9304d12ec38f9a9c2c43c3c80c2191ffc232ee6738eaba496
+size 572048820

language_model_model_layers_31.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa7a6849582268275ed44d2490acb31c5169bcb8a091c4c4dc1e804d568b8739
+size 572703808

language_model_model_layers_32.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:998a0890d850e11cba737244a48f3caa4632ac9041b1b44724c54995a58a9df7
+size 565932430

language_model_model_layers_33.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f41e839e0f409b222e5f469e29ded3105d451fd480a986d1f67ce350c54c13f
+size 573459544

language_model_model_layers_34.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93e548274791530ab06d86e74115e9aebbc6423b23b3ae1b8718b6b0c610820f
+size 567093312

language_model_model_layers_35.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb519ae031b8c6380111eb3ba3bf7f25ffa2f8ea82ced79dde0fbac382e8e0b9
+size 567542823

language_model_model_layers_36.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd991b547ae612673da5404c65fa041c5379e67eba60e69478232692c8c130d0
+size 573102208

language_model_model_layers_37.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ae2f7c9cb98c6ce8bebc6ccd331df4b1000aa4ace18557bdadad5e7ff6a438d
+size 566882054

language_model_model_layers_38.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ec73af5fded262305fb70b994c09b6aba9f4d5355907c2f75330250c5fd9dde
+size 568785902

language_model_model_layers_39.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a327bf17665d8823fa919faf99873e73a50714ad8bd5c9513df623ba3ecb1a53
+size 566043625

language_model_model_layers_4.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a5aad19bfff9aaeabecd29f4e8dbef4ead580799a8c811e877b9b0f7f9e3ddb
+size 568816050

language_model_model_layers_40.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32821353baff1a008a0a8a48a17debc11fe7224f3e9d5abdb00c69368f46b2d2
+size 565282420

language_model_model_layers_41.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:217e7f35bb2097f6d4761a5cc62b12bd58a3bf24e530529c795784e594d3d442
+size 568389315

language_model_model_layers_42.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8baba1e4a74add6f933d3df0f5918984a754625b6436e093439b638adfb51f68
+size 571082672

language_model_model_layers_43.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89ab0aec846b0e3cc7fc1693ea57089ade0e92c80890dd2fe5272c0dc13639f8
+size 566251477

language_model_model_layers_44.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16a2ed7c507fb32b278a8f27a54d9eef0b232a5ddb0359e4b93dba671214d920
+size 570801930

language_model_model_layers_45.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fbcd39a20db06337e93ca407faf98426ab7388988d9bfd2669789d4c829d825
+size 564268241

language_model_model_layers_46.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:325cfcdb1e75bdac54978565b0cbdd72e513b64ddee87a605df6f0bd6d85562d
+size 563799078

language_model_model_layers_47.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc76d919d486aab4eb14473e8a6793f610683f5c9c501a3a42a61cc151989c74
+size 570118066

language_model_model_layers_48.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c081c224ab8d0f5ea96fcba8f3f0f44c51c40b6ed2a632de7c6ded119d573106
+size 564951353

language_model_model_layers_49.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b7109c23d773f03a3c2295f2e18c89c01649957d4cc0ca94e19186d3e1eff3c
+size 563852471