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

config.json

@@ -0,0 +1,63 @@
+{
+  "name_or_path": "tclf90/GLM-4.7-GPTQ-Int4-Int8Mix",
+  "architectures": [
+    "Glm4MoeForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "pad_token_id": 151329,
+  "eos_token_id": [
+    151329,
+    151336,
+    151338
+  ],
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "partial_rotary_factor": 0.5,
+  "initializer_range": 0.02,
+  "intermediate_size": 12288,
+  "max_position_embeddings": 202752,
+  "model_type": "glm4_moe",
+  "moe_intermediate_size": 1536,
+  "norm_topk_prob": true,
+  "num_attention_heads": 96,
+  "n_group": 1,
+  "topk_group": 1,
+  "n_routed_experts": 160,
+  "n_shared_experts": 1,
+  "routed_scaling_factor": 2.5,
+  "num_experts_per_tok": 8,
+  "first_k_dense_replace": 3,
+  "num_hidden_layers": 92,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "num_nextn_predict_layers": 1,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.54.0",
+  "use_cache": true,
+  "use_qk_norm": true,
+  "vocab_size": 151552,
+  "quantization_config": {
+    "quant_method": "gptq",
+    "bits": 4,
+    "group_size": 128,
+    "sym": true,
+    "desc_act": false,
+    "dynamic": {
+      "-:model.embed_tokens.weight": {},
+      "-:.*shared_experts": {},
+      "-:.*shared_head": {},
+      "-:lm_head.weight": {},
+      "+:model[.]layers[.].*[.]self_attn[.].*": {
+        "bits": 8
+      },
+      "+:model[.]layers[.]([0-3]|89|9[0-2])[.].*": {
+        "bits": 8
+      }
+    }
+  }
+}

configuration.json

@@ -0,0 +1 @@
+{"framework":"Pytorch","task":"text-generation"}

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    151329,
+    151336,
+    151338
+  ],
+  "pad_token_id": 151329,
+  "temperature": 1.0,
+  "transformers_version": "4.56.2"
+}

model-00006-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e02c584a4b98a81dbbcd6fc0957c8a10ca16fb87497e79db05120f8d63f7a9d0
+size 2998138016

model-00009-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09e711b072c2086bdd9d4a1ff31c21ea92943a70d1e9d61cb913cc50ec78a9ec
+size 2998336728

model-00012-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95643a168a7184a7978a5d3f66ad690b5f8befd9786db3a8354d364c52ff6725
+size 2998336368

model-00014-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e4725ae78c8d157200baeb1a15aec1529f04cdc8fc75f7eb1f078b968361d1e
+size 2998336792

model-00019-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3878031ab5a8361297b04028dd1fd6c295da90d9ec3fc41e356156dcebb465f4
+size 2997202784

model-00025-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1422d3f6c4acc4a9f234ccc5c2022662384e748517eb0b513f97839ab81c05c4
+size 2998336256

model-00026-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea5d1c68e87429c3a47ab5523c5f20e9ab7c614e25c7ec3c621c32e20afb02b9
+size 2998140120

model-00031-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c39fcd1e752f527a673da82fb847337842804c8e10792ea4d8ce76c8fa1d9801
+size 2998140048

model-00040-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df231db86782e1c2e921e5b763bc2561491f279073c5132577f6519b604ac1f3
+size 2998322272

model-00044-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecf9c0dbd4c315838139e66085a82857262e73c473d3ea97e2de34cf6d5af11a
+size 2998140120

model-00047-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9845c4ee8bdbea110919de982567c6adf5e803cb227ecd38fdaf769e3653ec2
+size 2998140472

model-00054-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e02dcd22bc961f2188a33648cc78105cdee5dce5ba14a31fcac75b012bdbf2f5
+size 2966449120

model-00057-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5101f4f671a6ca9e8caecffac4a402ecb662aee923d2a315f1f77d9d0441354
+size 2998140224

model-00060-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40b4d3302f7b9868fbe560feffd54d6a69f7f44b63d94679495d05675066dd31
+size 2998140640

model-00061-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a72798c42e76e3ae460ca062f6818dd4b58862eecdd280c8d694f912b160bb2d
+size 2998336256

model-00065-of-00071.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7919e4ae240df7448aeb4d3d2d6f6ad95a684cd35ed875263f244ec01fcd37bf
+size 2995057976

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9340665016419c825c4bdabbcc9acc43b7ca2c68ce142724afa829abb1be5efd
+size 19970699