Upload SwitchTransformersForConditionalGeneration

config.json

@@ -0,0 +1,44 @@
+{
+  "_name_or_path": "hf-internal-testing/tiny-random-SwitchTransformersForConditionalGeneration",
+  "add_router_probs": false,
+  "architectures": [
+    "SwitchTransformersForConditionalGeneration"
+  ],
+  "bos_token_id": 0,
+  "d_ff": 37,
+  "d_kv": 8,
+  "d_model": 32,
+  "decoder_sparse_step": 1,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "relu",
+  "dropout_rate": 0.1,
+  "encoder_sparse_step": 1,
+  "eos_token_id": 1,
+  "expert_capacity": 100,
+  "feed_forward_proj": "relu",
+  "initializer_factor": 0.002,
+  "is_encoder_decoder": true,
+  "is_gated_act": false,
+  "layer_norm_epsilon": 1e-06,
+  "model_type": "switch_transformers",
+  "num_decoder_layers": 5,
+  "num_experts": 8,
+  "num_heads": 4,
+  "num_layers": 5,
+  "num_sparse_decoder_layers": 3,
+  "num_sparse_encoder_layers": 3,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 8,
+  "router_aux_loss_coef": 0.001,
+  "router_bias": false,
+  "router_dtype": "float32",
+  "router_ignore_padding_tokens": false,
+  "router_jitter_noise": 0.0,
+  "router_type": "tokens_masked",
+  "router_z_loss_coef": 0.001,
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.1",
+  "use_cache": true,
+  "vocab_size": 1302
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3e1f2bd2d1c11018025aaef0d05fb85c1616e943fa6c29e2da4f252da593122
+size 546501