Add files using upload-large-folder tool

README.md

@@ -2,8 +2,6 @@
 library_name: diffusers
 tags:
 - pruna-ai
-base_model:
-- black-forest-labs/FLUX.1-schnell
 ---
 
 # Model Card for PrunaAI/FLUX.1-schnell-smashed
@@ -41,21 +39,24 @@ The compression configuration of the model is stored in the `smash_config.json`
     "batcher": null,
     "cacher": "fora",
     "compiler": "torch_compile",
-    "factorizer": "qkv_diffusers",
+    "factorizer": null,
     "pruner": null,
-    "quantizer": null,
-    "fora_interval": 2,
+    "quantizer": "torchao",
+    "fora_interval": 3,
     "fora_start_step": 2,
     "torch_compile_backend": "inductor",
     "torch_compile_dynamic": null,
     "torch_compile_fullgraph": true,
     "torch_compile_make_portable": false,
     "torch_compile_max_kv_cache_size": 400,
-    "torch_compile_mode": "default",
+    "torch_compile_mode": "max-autotune-no-cudagraphs",
     "torch_compile_seqlen_manual_cuda_graph": 100,
     "torch_compile_target": "model",
+    "torchao_excluded_modules": "norm+embedding",
+    "torchao_quant_type": "int8dq",
     "batch_size": 1,
-    "device": "cuda",
+    "device": "cpu",
+    "device_map": null,
     "save_fns": [
         "save_before_apply",
         "save_before_apply"
@@ -64,9 +65,9 @@ The compression configuration of the model is stored in the `smash_config.json`
         "diffusers"
     ],
     "reapply_after_load": {
-        "factorizer": "qkv_diffusers",
+        "factorizer": null,
         "pruner": null,
-        "quantizer": null,
+        "quantizer": "torchao",
         "cacher": "fora",
         "compiler": "torch_compile",
         "batcher": null

model_index.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "FluxPipeline",
-  "_diffusers_version": "0.33.1",
+  "_diffusers_version": "0.34.0",
   "_name_or_path": "black-forest-labs/FLUX.1-schnell",
   "feature_extractor": [
     null,

scheduler/scheduler_config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "FlowMatchEulerDiscreteScheduler",
-  "_diffusers_version": "0.33.1",
+  "_diffusers_version": "0.34.0",
   "base_image_seq_len": 256,
   "base_shift": 0.5,
   "invert_sigmas": false,
@@ -9,6 +9,7 @@
   "num_train_timesteps": 1000,
   "shift": 1.0,
   "shift_terminal": null,
+  "stochastic_sampling": false,
   "time_shift_type": "exponential",
   "use_beta_sigmas": false,
   "use_dynamic_shifting": false,

smash_config.json

@@ -2,21 +2,24 @@
     "batcher": null,
     "cacher": "fora",
     "compiler": "torch_compile",
-    "factorizer": "qkv_diffusers",
+    "factorizer": null,
     "pruner": null,
-    "quantizer": null,
-    "fora_interval": 2,
+    "quantizer": "torchao",
+    "fora_interval": 3,
     "fora_start_step": 2,
     "torch_compile_backend": "inductor",
     "torch_compile_dynamic": null,
     "torch_compile_fullgraph": true,
     "torch_compile_make_portable": false,
     "torch_compile_max_kv_cache_size": 400,
-    "torch_compile_mode": "default",
+    "torch_compile_mode": "max-autotune-no-cudagraphs",
     "torch_compile_seqlen_manual_cuda_graph": 100,
     "torch_compile_target": "model",
+    "torchao_excluded_modules": "norm+embedding",
+    "torchao_quant_type": "int8dq",
     "batch_size": 1,
-    "device": "cuda",
+    "device": "cpu",
+    "device_map": null,
     "save_fns": [
         "save_before_apply",
         "save_before_apply"
@@ -25,9 +28,9 @@
         "diffusers"
     ],
     "reapply_after_load": {
-        "factorizer": "qkv_diffusers",
+        "factorizer": null,
         "pruner": null,
-        "quantizer": null,
+        "quantizer": "torchao",
         "cacher": "fora",
         "compiler": "torch_compile",
         "batcher": null

text_encoder/config.json

@@ -19,6 +19,6 @@
   "pad_token_id": 1,
   "projection_dim": 768,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.52.4",
+  "transformers_version": "4.53.1",
   "vocab_size": 49408
 }

text_encoder_2/config.json

@@ -25,7 +25,7 @@
   "relative_attention_num_buckets": 32,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.52.4",
+  "transformers_version": "4.53.1",
   "use_cache": true,
   "vocab_size": 32128
 }

text_encoder_2/model.safetensors.index.json

@@ -1,5 +1,6 @@
 {
   "metadata": {
+    "total_parameters": 4762310656,
     "total_size": 9524621312
   },
   "weight_map": {

transformer/config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "FluxTransformer2DModel",
-  "_diffusers_version": "0.33.1",
+  "_diffusers_version": "0.34.0",
   "_name_or_path": "/root/.cache/huggingface/hub/models--black-forest-labs--FLUX.1-schnell/snapshots/741f7c3ce8b383c54771c7003378a50191e9efe9/transformer",
   "attention_head_dim": 128,
   "axes_dims_rope": [

vae/config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "AutoencoderKL",
-  "_diffusers_version": "0.33.1",
+  "_diffusers_version": "0.34.0",
   "_name_or_path": "/root/.cache/huggingface/hub/models--black-forest-labs--FLUX.1-schnell/snapshots/741f7c3ce8b383c54771c7003378a50191e9efe9/vae",
   "act_fn": "silu",
   "block_out_channels": [