:tada: uploaded baseline model

README.md

@@ -1,3 +1,72 @@
 ---
 license: apache-2.0
+tags:
+- jax
+- safetensors
 ---
+
+# Baseline PerceptNet
+
+## Model Description
+
+## How to use it
+
+### Install the model's package from source:
+```
+git clone https://github.com/Jorgvt/paramperceptnet.git
+cd paramperceptnet
+pip install -e .
+```
+
+### 1.Import necessary libraries:
+
+```
+import json
+
+from huggingface_hub import hf_hub_download
+import flax
+import orbax.checkpoint
+from ml_collections import ConfigDict
+
+from paramperceptnet.models import Baseline as PerceptNet
+```
+
+### 2.Download the configuration
+
+```
+config_path = hf_hub_download(repo_id="Jorgvt/ppnet-baseline",
+                              filename="config.json")
+with open(config_path, "r") as f:
+    config = ConfigDict(json.load(f))
+```
+
+### 3. Download the weights
+
+#### 3.1. Using `safetensors`
+
+```
+from safetensors.flax import load_file
+
+weights_path = hf_hub_download(repo_id="Jorgvt/ppnet-baseline",
+                               filename="weights.safetensors")
+variables = load_file(weights_path)
+variables = flax.traverse_util.unflatten_dict(variables, sep=".")
+params = variables["params"]
+```
+
+#### 3.2. Using `mgspack`
+```
+weights_path = hf_hub_download(repo_id="Jorgvt/ppnet-fully-trained",
+                               filename="weights.msgpack")
+with open(weights_path, "rb") as f:
+    variables = orbax.checkpoint.msgpack_utils.msgpack_restore(f.read())
+variables = jax.tree_util.tree_map(lambda x: jnp.array(x), variables)
+params = variables["params"]
+```
+
+### 4. Use the model
+
+```
+from jax import numpy as jnp
+pred = model.apply({"params": params}, jnp.ones((1,384,512,3)))
+```

config.json

@@ -0,0 +1 @@
+{"A_GABOR": true, "A_GDNSPATIOFREQORIENT": true, "BATCH_SIZE": 64, "CS_KERNEL_SIZE": 5, "END_LR": 0.005, "EPOCHS": 500, "GABOR_KERNEL_SIZE": 5, "GDNFINAL_KERNEL_SIZE": 1, "GDNGAUSSIAN_KERNEL_SIZE": 1, "GDN_CLIPPING": true, "INITIAL_LR": 0.01, "INIT_GABOR": false, "INIT_JH": false, "LEARNING_RATE": 0.003, "NORMALIZE_ENERGY": true, "NORMALIZE_PROB": false, "PEAK_LR": 0.04, "SEED": 42, "TRAIN_CS": false, "TRAIN_GABOR": false, "TRAIN_GDNCOLOR": false, "TRAIN_GDNGAMMA": false, "TRAIN_GDNGAUSSIAN": false, "TRAIN_JH": false, "TRAIN_ONLY_LAST_GDN": true, "USE_BIAS": false, "USE_GAMMA": false, "WARMUP_EPOCHS": 15, "ZERO_MEAN": true}

weights.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77fd6911c18eb9813a955b568682809855201e4396889f34774920047f721e5d
+size 30395788

weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e0876beebc3fe5bff1ae63f484ce1ab25a90dbf624757f4962c888ef9d808a8
+size 30396424