Upload Idefics2ForConditionalGeneration

config.json

@@ -1,6 +1,6 @@
 {
   "architectures": [
-    "Idefics2Model"
+    "Idefics2ForConditionalGeneration"
   ],
   "image_token_id": 32001,
   "model_type": "idefics2",
@@ -95,7 +95,7 @@
     "forced_bos_token_id": null,
     "forced_eos_token_id": null,
     "hidden_act": "silu",
-    "hidden_size": 16,
+    "hidden_size": 32,
     "id2label": {
       "0": "LABEL_0",
       "1": "LABEL_1"
@@ -110,11 +110,11 @@
     },
     "length_penalty": 1.0,
     "max_length": 20,
-    "max_position_embeddings": 131072,
+    "max_position_embeddings": 32768,
     "min_length": 0,
     "model_type": "mistral",
     "no_repeat_ngram_size": 0,
-    "num_attention_heads": 2,
+    "num_attention_heads": 8,
     "num_beam_groups": 1,
     "num_beams": 1,
     "num_hidden_layers": 2,
@@ -176,7 +176,7 @@
     "forced_bos_token_id": null,
     "forced_eos_token_id": null,
     "hidden_act": "gelu_pytorch_tanh",
-    "hidden_size": 16,
+    "hidden_size": 256,
     "id2label": {
       "0": "LABEL_0",
       "1": "LABEL_1"
@@ -196,11 +196,11 @@
     "min_length": 0,
     "model_type": "idefics2",
     "no_repeat_ngram_size": 0,
-    "num_attention_heads": 2,
+    "num_attention_heads": 4,
     "num_beam_groups": 1,
     "num_beams": 1,
     "num_channels": 3,
-    "num_hidden_layers": 12,
+    "num_hidden_layers": 2,
     "num_return_sequences": 1,
     "output_attentions": false,
     "output_hidden_states": false,

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.40.2"
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9fd6f8069ebced264f5aeb2d5ad3a44d127bcbcbae721579b925ae2cf6cd588b
-size 3271296
+oid sha256:b6de3469bdf7124dc1615286bdbcec09d82a22314fba204fafab573ba132199a
+size 15634952